The ankylosaurid dinosaurs of the Upper Cretaceous Baruungoyot and Nemegt formations of Mongolia

Abstract

The discovery of a new ankylosaurid skull with some unusual features from the Baruungoyot Formation of Mongolia prompted a systematic review of ankylosaurid specimens from the Baruungoyot and Nemegt formations. Dyoplosaurus giganteus was found to possess no diagnostic features and is regarded as a nomen dubium. The holotype of Tarchia kielanae (previously synonymized with Tarchia gigantea) has one autapomorphy, an accessory postorbital ossification with surrounding furrow, and Tar. kielanae is here considered a valid species, making the combination Tar. gigantea unnecessary. An accessory postorbital ossification is also found in the holotype of Minotaurasaurus ramachandrani, and this species is here considered a junior synonym of Tar. kielanae. The newly described skull from the Baruungoyot Formation forms the holotype of a new genus and species, Zaraapelta nomadis gen. et sp. nov., diagnosed by unusual bilayered ornamentation on the squamosal horn and extensive postocular ornamentation. Two distinct tail club handle morphotypes are present in the Nemegt Formation and probably represent two different species. However, it is impossible to assign either tail club morphotype to the single valid species from the formation, Saichania chulsanensis, because of a lack of overlapping material. A revised phylogenetic analysis including newly identified characters found Zaraapelta nomadis to be most closely related to Tar. kielanae.

Full text

This is the accepted version of the following article:
Arbour VM, Currie PJ, Badamgarav D. 2014. The ankylosaurid dinosaurs of the Upper
Cretaceous Baruungoyot and Nemegt formations of Mongolia. Zoological Journal of the
Linnean Society 172:631-652.
The published final version can be found at:
http://onlinelibrary.wiley.com/doi/10.1111/zoj.12185/abstract

ABSTRACT
The discovery of a new ankylosaurid skull with some unusual features from the Baruungoyot
Formation of Mongolia prompted a systematic review of ankylosaurid specimens from the Baruungoyot
and Nemegt formations. Dyoplosaurus giganteus was found to possess no diagnostic features and is
regarded as a nomen dubium. The holotype of Tarchia kielanae, (previously synonymized with Tarchia
gigantea) has one autapomorphy, an accessory postorbital ossification with surrounding furrow, and
Tarchia kielanae is here considered a valid species, making the combination Tarchia gigantea
unnecessary. An accessory postorbital ossification is also found in the holotype of Minotaurasaurus
ramachandrani, and this species is here considered a junior synonym of Tarchia kielanae. The newly
described skull from the Baruungoyot Formation forms the holotype of a new genus and species,
Zaraapelta nomadis, diagnosed by unusual bi-layered ornamentation on the squamosal horn and
extensive postocular ornamentation. Two distinct tail club handle morphotypes are present in the
Nemegt Formation and probably represent two different species. However, it is impossible to assign
either tail club morphotype to the single valid species from the formation, Saichania chulsanensis,
because of a lack of overlapping material. A revised phylogenetic analysis including newly identified
characters found Zaraapelta nomadis to be most closely related to Tarchia kielanae.
Keywords: Ankylosauria-Ankylosauridae-Campanian-Dinosauria- Gobi Desert-Maastrichtian
INTRODUCTION
The Upper Cretaceous Baruungoyot and Nemegt formations of the Gobi Desert, Mongolia (Fig.
1), have produced several ankylosaurid specimens with distinctive bulbous, pyramidal cranial
ornamentation. Three ankylosaurid taxa from the Baruungoyot and Nemegt formations were named
during the 20th century: Dyoplosaurus giganteus Maleev, 1956, Saichania chulsanensis Maryańska, 1977,
and Tarchia kielanae Maryańska, 1977. The holotype of a fourth taxon, Minotaurasaurus ramachandrani
Miles & Miles, 2009, was purchased from the Tucson Gem, Mineral and Fossil Showcase (Arizona, USA)
without provenance data, but it has been suggested that the skull was also collected in Mongolia
(Dalton, 2009). Dyoplosaurus giganteus was reassigned to the genus Tarchia (as Tarchia gigantea) by
Tumanova (1977). The holotype skull of Tarchia kielanae, ZPAL MgD I/111, is poorly preserved;
therefore, most comparisons and phylogenetic analyses have used PIN 3142/250, a referred specimen,
to represent Tarchia gigantea (Vickaryous, Maryańska, & Weishampel, 2004; Miles & Miles 2009;
Thompson et al., 2012). Tarchia gigantea and Saichania chulsanensis have been recovered as sister taxa
in subsequent phylogenetic analyses (Carpenter, 2001; Vickaryous et al., 2004; Thompson et al., 2012).
The discovery of a new skull with some unusual features (MPC D100/1338) from the
Baruungoyot Formation at Hermiin Tsav prompted the following reassessment of the morphology and
taxonomic assignments of previously described ankylosaurid specimens from Mongolia. New cranial
characters are identified, and a revised phylogenetic analysis is conducted in order to determine the
evolutionary relationships within derived ankylosaurids.

INSTITUTIONAL ABBREVIATIONS
AMNH, American Museum of Natural History, New York, New York, USA; HBV, Earth Sciences
Museum of Shijiazhuang University of Economics, Hebei, China; IMM, Inner Mongolia Museum, Hohhot,
China; INBR, Victor Valley Museum, California, USA; MPC, Paleontological Center, Mongolian Academy
of Sciences, Ulaanbaatar, Mongolia (MPC KID refers to Korea-Mongolia International Dinosaur Project
field numbers); NHMUK, Natural History Museum, London, United Kingdom; PIN, Paleontological
Institute, Russian Academy of Sciences, Moscow, Russia; ZPAL, Zoological Institute of Paleobiology,
Polish Academy of Sciences, Warsaw, Poland.
ANATOMICAL ABBREVIATIONS
V, opening for trigeminal nerve; VI, opening for abducens nerve; VII, opening for facial nerve;
XII, opening for hypoglossal nerve; acc po, accessory postorbital ossification; alv, alveolar ridge; ang,
angular; asca, anterior supraorbital caputegulum; aso, anterior supraorbital; bas, basioccipital; bpt,
basipterygoid process; bs, basisphenoid; c, centrum; cor, coronoid; cp, crista prootica; dent, dentary;
dpf, descending process of frontal; ee, ectethmoid; fm, foramen magnum; fo, fenestra ovalis (fenestra
vestibularis); fr, frontal; frca, frontal caputegulum; furrow, furrow in postorbital; ha, haemal arch; ic,
internal carotid artery; inca, internarial caputegulum; j, jugal; jf, jugular foramen; kn, tail club knob; lac,
lacrimal; lac inc, lacrimal incisure; laca, lacrimal caputegulum; loca, loreal caputegulum; ls,
laterosphenoid; ltf, laterotemporal fenestra; maca, mandibular caputegulum; mand, mandible; msca,
middle supraorbital caputegulum; mso, middle supraorbital; mx, maxilla; nar, narial opening; nar apt,
narial aperture; nas, nasal; nas vest, nasal vestibule; nasca, nasal caputegulum; nc, neural canal; ns,
nasal septum; nsp, neural spine; np, nasal passage; nuca, nuchal caputegulum; nuch, nuchal crest; oc,
occipital condyle; occ, occiput; ocv, orbitocerebral vein; orb, orbit; orbs, orbitosphenoid; os, osteoderm;
ot, ossified tendon; par, parietal; parocc, paroccipital process; path, pathology; pmx, premaxilla; pmx
orn, premaxillary ornamentation; prezyg, prezygapophysis; po, postorbital; poca, postocular
caputegulum; preart, prearticular; predent, predentary; prf, prefrontal; prfca, prefrontal caputegulum;
pro alv, alveolar border; pro retro, retroarticular process; ps, parasphenoid; psca, postorbital
supraorbital caputegulum; pso, posterior supraorbital; pt, pterygoid; ptv, interpterygoid vacuity; q,
quadrate; qh, quadrate head; qj, quadratojugal; qjh, quadratojugal horn; snca, supranarial caputegulum;
socc, supraoccipital; sp, sinus of pituitary; sq, squamosal; sqh, squamosal horn; sulc predent, predentary
sulcus; sur, surangular; tp, transverse process; u, ungual; v, vomer.
MATERIALS AND METHODS
Original specimens were examined, measured, and photographed where possible,
supplemented by observation of casts and the literature (Appendix 1). Measurements were taken using
digital calipers and a measuring tape.
Ankylosaur cranial ornamentation is useful for differentiating ankylosaur taxa, and in this paper
the term 'caputegulum' (Latin: "skull tile"), originally coined by Blows (2001), is used to refer to the
polygonal sculpturing on an ankylosaur skull. Although caputegulum originally referred only to flat
cranial sculpturing (Blows, 2001), it can also be applied to the bulbous, discrete cranial ornamentation
found in some ankylosaurids. Arbour & Currie (2013a) use caputegulum with a location modifier (e.g.
prefrontal caputegulum, supranarial caputegulum) to compare ornamentation patterns in specimens

referred to Euoplocephalus tutus (Lambe, 1902) and these terms are used here and modified where
necessary. Mongolian place names and stratigraphic units follow the spelling conventions outlined in
Benton (2000).
The phylogenetic relationships of the Mongolian taxa within the Ankylosauridae were assessed
using TNT v1.1 (Goloboff, Farris & Nixon, 2008). The data matrix was assembled in Mesquite version
2.72 (Maddison & Maddison, 2011) and included 160 characters for 20 species. Outgroup taxa included
Lesothosaurus diagnosticus Galton, 1978 (a basal ornithischian), Scelidosaurus harrisonii Owen, 1861(a
basal thyreophoran), Stegosaurus spp. (a stegosaur), the nodosaurid ankylosaurs Panoplosaurus mirus
Lambe, 1919, and Pawpawsaurus campbelli Lee, 1996, and the basal ankylosaur Gastonia burgei
Kirkland, 1998. New ankylosaurid cranial characters have been identified and incorporated into a revised
data matrix modified from Arbour & Currie (2013a). Because several important specimens have been
reassigned at the species level, character codings for Tarchia are significantly different than those in
previous analyses. Characters were treated as unordered and of equal weight. The parsimony analysis
conducted in TNT used the Traditional Search option with one random seed and 1000 replicates of
Wagner trees and the tree bisection reconnection (TBR) swapping algorithm.
SYSTEMATIC PALAEONTOLOGY
DINOSAURIA OWEN, 1842
ORNITHISCHIA SEELEY, 1887
THYREOPHORA NOPCSA, 1915
ANKYLOSAURIA OSBORN, 1923
ANKYLOSAURIDAE BROWN, 1908
ANKYLOSAURINAE NOPCSA, 1929
DYOPLOSAURUS GIGANTEUS MALEEV, 1956
Holotype: PIN 551/29, series of caudal vertebrae, metatarsals, phalanges, osteoderms; also includes a
partial tail club knob not described or figured by Maleev (1956) or subsequent authors.
Holotype locality and stratigraphy: Nemegt, Mongolia; Nemegt Formation (Upper Campanian – Lower
Maastrichtian; Jerzykiewicz, 2000)
Previous diagnosis: From Maleev (1956): Anterior caudal vertebrae short, high, amphicoelous; chevrons
massive, fused to vertebra; distal caudal vertebrae long, low; metatarsal bones short, wide; unguals
thick, hoof-shaped; osteoderms sharp, thin-walled, with numerous pits and channels on the external
surface.
Status: Nomen dubium.
Discussion: Maleev (1956) assigned PIN 551/29 (Fig. 2) to the genus Dyoplosaurus based on the
similarity of the free caudal and handle caudal vertebrae to those of the North American taxon
Dyoplosaurus acutosquameus Parks, 1924, but erected the new species Dyoplosaurus giganteus based
on the greater size of PIN 551/29. PIN 551/29 is larger than ROM 784 (D. acutosquameus): the largest
free caudal vertebra in ROM 784 is 63 mm high, whereas the largest preserved free caudal in PIN 551/29
is 126 mm high. However, because differences in size can result from ontogenetic or individual variation

in addition to taxonomic variation, differentiating Dyoplosaurus giganteus from Dyoplosaurus
acutosquameus based only on size differences is insufficient.
The preserved elements of PIN 551/29 do not differ significantly from any other Late Cretaceous
ankylosaurine ankylosaurids, except for Ankylosaurus magniventris Brown, 1908, and one specimen also
from the Nemegt Formation, ZPAL MgD I/113. In PIN 551/29, the handle caudal vertebrae have V-
shaped neural spines in which the prezygapophyses diverge at an angle of about 22-26° (Fig. 2E). This is
the typical condition for most ankylosaurids, including Euoplocephalus tutus, Pinacosaurus grangeri
Gilmore, 1933, and Talarurus plicatospineus Maleev, 1952 (Arbour, Burns & Sissons, 2009). Ankylosaurus
magniventris has U-shaped neural spines in which the prezygapophyses diverge at an angle of about 60°
(Arbour et al., 2009). ZPAL MgD I/113, a nearly complete tail from the Nemegt Formation, has handle
vertebrae that have an intermediate morphology between V-shaped (e.g. Euoplocephalus tutus) and U-
shaped (Ankylosaurus magniventris), with prezygapophyses diverging at an angle of about 37° (Arbour
et al., 2009; Fig. 2F). Although the handle vertebrae of PIN 551/29 differ from those of ZPAL MgD I/113
and specimens referred to Ankylosaurus magniventris, they are indistinguishable from the handle
vertebrae of all other ankylosaurines.
The proposed diagnostic features of D. giganteus (short, high, amphicoelous caudal vertebrae
with coossified massive chevrons; long, low distal caudal vertebrae; short, wide metatarsal bones; thick,
hoof-shaped unguals; and sharp, thin-walled osteoderms with numerous pits and channels) are present
in all ankylosaurines where these features are preserved, and so cannot be considered diagnostic of
Dyoplosaurus or D. giganteus. PIN 551/29 has no autapomorphies, nor a unique combination of
characters that differentiates this specimen from other ankylosaurines. As such, Dyoplosaurus giganteus
must be regarded as a nomen dubium.
TARCHIA KIELANAE MARYAŃSKA, 1977
= Minotaurasaurus ramachandrani Miles and Miles, 2009
Holotype: ZPAL MgD I/111, posterior part of skull roof, braincase, and partial occiput.
Holotype locality and stratigraphy: Khulsan, Mongolia; Baruungoyot Formation (Mid-Upper Campanian,
Jerzykiewicz, 2000)
Referred specimen: INBR21004, complete skull, mandibles, and predentary (holotype of
Minotaurasaurus ramachandrani; provenance unknown)
Revised Differential Diagnosis: Ankylosaurine ankylosaurid with bulbous frontonasal cranial
ornamentation. Uniquely among ankylosaurines, Tarchia kielanae has a discrete postorbital ossification
separate from, but adjacent to, the squamosal horn, surrounded by a furrow; four bulbous internarial
caputegulae; smooth and widely flaring supranarial caputegulae; a triangular region of rugose
ornamentation with discrete edge on the premaxilla ventral to nasal vestibule; and pterygoid body more
horizontally oriented than in other ankylosaurids.
Previous diagnoses: From Maryańska (1977), for Tarchia kielanae: Orbits not completely closed;
exoccipital high and short, perpendicular to skull roof; occipital condyle directed posteroventrally;
foramen magnum higher than wide; braincase tall; occipital condyle and occiput partly visible in dorsal
view; one cranial nerve opening situated posterior to foramen ovale. From Miles & Miles (2009), for
Minotaurasaurus ramachandrani: Large, horizontally elliptical external nares situated terminally;
external nares rimmed laterally and posteriorly by well-developed osteoderm; anteriorly rimmed by

thin, triangular osteoderm fused on premaxilla; foramina for premaxillary and maxillary sinuses housed
within external nares; premaxillary part of snout broad; occipital condyle poorly developed (as in
Saichania chulsanensis); occipital condyle directed ventrally; exoccipitals low and separated from skull
roof by gap; dorsal part of exoccipitals near supraoccipital curved anterodorsally; quadrate nearly
vertical; quadrate head not fused to paroccipital process; skull roof not overhanging occiput; maxillary
shelf well-developed and wide to below middle of orbit; premaxilla forms anterior rim of palatal vacuity
and separating maxillae from vomer (as in Pinacosaurus); premaxillary beak wider than distance
between posterior maxillary tooth rows; pterygoid body horizontal (not vertical as in Saichania
chulsanensis, Tarchia kielanae, most ankylosaurids); teeth similar to Pinacosaurus with weakly
developed cingulum.
Status: Valid
Discussion: Tumanova (1977), noting similarities between the newly-collected specimen PIN 3142/250,
ZPAL MgD I/111 (holotype of Tarchia kielanae), and PIN 551/29 (holotype of Dyoplosaurus giganteus),
reassigned Dyoplosaurus giganteus to Tarchia to form the new combination Tarchia gigantea; Tarchia
kielanae was retained as a separate species. Although not explicitly stated, it is likely that Tarchia was
favoured as the generic name over Dyoplosaurus, which has priority, because at the time Dyoplosaurus
acutosquameus was considered a junior synonym of Euoplocephalus tutus (Coombs, 1971, 1978), and
PIN 3142/150 was clearly not referable to Euoplocephalus tutus. In a comprehensive review of
ankylosaurs from Mongolia, Tumanova (1987) retained Tarchia kielanae as a separate species, but noted
that Tarchia kielanae and Tarchia gigantea could not be easily compared due to the fragmentary nature
of ZPAL MgD I/111 (Fig. 3B); she suggested that Tarchia kielanae may be a synonym of Tarchia gigantea.
Tarchia kielanae was later regarded as a junior synonym of Tarchia gigantea by Coombs & Maryańska
(1990) and subsequent authors. Arbour et al. (2009) did not address the validity of Dyoplosaurus
giganteus in their reappraisal of Dyoplosaurus acutosquameus. When Tarchia is discussed in the
literature, comparisons are usually made with the more complete specimen PIN 3142/150 (Figs. 3A, 4E,
5A) rather than the holotype skull ZPAL MgD I/111. Although fragmentary, there are several important
differences between the skulls of PIN 3142/150 and ZPAL MgD I/111, which indicate that these
specimens do not represent the same taxon. As such, Tarchia kielanae is here regarded as a valid taxon,
but PIN 3142/150 is not referable to Tarchia kielanae.
Several of the diagnostic characters proposed for Tarchia kielanae by Maryańska (1977) have a
broad distribution within Ankylosauridae. A high and short paroccipital process oriented perpendicular
to the skull roof, and posteroventrally directed occipital condyle, are present in all ankylosaurids, e.g.
Ankylosaurus magniventris (Carpenter 2004), Anodontosaurus lambei and Euoplocephalus tutus (Arbour
& Currie 2013a; Vickaryous & Russell 2003), Pinacosaurus grangeri (Maryańska 1977), and Tsagantegia
longicranialis (Tumanova 1993). The orbits are not bordered by a fully developed bony wall anteriorly
like in Euoplocephalus tutus (Miyashita et al. 2011) or Saichania chulsanensis (Maryańska 1977), but the
new taxon described in this paper also lacks this feature. Other diagnostic characters are too vague to
assess, such as a 'tall braincase'. The foramen magnum is marginally taller than it is wide, but the
dimensions of the foramen magnum vary in specimens of Anodontosaurus lambei and Euoplocephalus
tutus (Arbour & Currie 2013a).
ZPAL MgD I/111 could not be located at ZPAL in October 2009, so the black and white
photographs in Maryańska (1977) were used for comparisons. Much of the skull of ZPAL MgD I/111 is

missing, and it is difficult to tell exactly which edges are broken, versus which edges are complete, in the
figures in Maryańska (1977). The ventral view of ZPAL MgD I/111 provided by Maryańska (1977:Pl. 27)
shows that the distal (lateral) end of the right paroccipital process is missing. The paroccipital processes
are fused to the parietals and squamosals, and so the absence of the paroccipital in this region indicates
that there is substantial damage to the posterior edge of the skull roof. Nevertheless, there are two
notable features in the preserved part of ZPAL MgD I/111 that are unique: the occiput is visible in dorsal
view, and the preserved portion of the 'squamosal' horn is surrounded anteriorly and laterally by a
pronounced furrow (Fig. 3B). The posterior margin of the skull roof is broken in ZPAL MgD I/111, so it is
difficult to determine with certainty if the occiput was truly visible in dorsal view; however, the occipital
condyle is also damaged, and may have extended further posteriorly originally.
The furrow around the 'squamosal' horn is present in only one other known ankylosaur
specimen, the holotype of Minotaurasaurus ramachandrani; comparison (Fig. 3C) shows that the furrow
in ZPAL MgD I/111 corresponds to the unusual discrete postorbital ossification in INBR21004, rather
than the squamosal horn proper. In contrast, no groove near the squamosal horn is present in PIN
3142/250 (referred Tarchia gigantea specimen, Fig. 3A), there is no evidence for a distinct postorbital
ossification, and the occipital condyle is not visible in dorsal view. Miles & Miles (2009), in their
description of Minotaurasaurus ramachandrani, do not explicitly state whether or not they are
comparing Minotaurasaurus to ZPAL MgD I/111 (Tarchia kielanae) or PIN 3142/250 (Tarchia gigantea);
no specimen numbers, explicit citations to original literature, or species names are provided in their
comparisons. Maryańska (1977) is listed as a reference in Miles & Miles (2009), but the only in-text
citation is in reference to the general morphology of ankylosaurid mandibles. The comparison between
Minotaurasaurus ramachandrani and Tarchia by Miles & Miles (2009) includes references to features
that are not preserved in ZPAL MgD I/111 (Tarchia kielanae), like the premaxilla, maxilla, the shape of
the orbits, pterygoid body, quadrate, quadratojugal horns, distal ends of the paroccipital processes, and
mandibles (including predentary). For this reason, it appears most likely that comparisons between
Minotaurasaurus ramachandrani and Tarchia by Miles & Miles (2009) are primarily comparisons
between INBR21004 and PIN 3142/250 (specimen previously referred to Tarchia gigantea) and not
INBR21004 and ZPAL MgD I/111 (Tarchia kielanae). The preserved portions of ZPAL MgD I/111 differ in
only one respect from the comparable portions of INBR21004: the supraoccipital does not appear to be
coossified to the parietals in INBR21004, but is in ZPAL MgD I/111. This could represent either an
ontogenetic or taxonomic difference between the two specimens. The accessory postorbital ossification
surrounded by a distinct furrow is present in only two ankylosaurid specimens, INBR21004 and ZPAL
MgD I/114, and because the two skulls share no other significant differences, we consider
Minotaurasaurus ramachandrani a junior synonym of Tarchia kielanae.
Several of the diagnostic characters for Minotaurasaurus ramachandrani by Miles & Miles
(2009) are retained with moderate revisions here as part of the revised diagnosis for Tarchia kielanae.
Other diagnostic characters proposed by Miles & Miles (2009) are found in multiple ankylosaurid taxa:
large, horizontally elliptical, terminally situated external nares, a broad premaxillary region of the snout,
a well-developed maxillary shelf, the orientation of the quadrate, and a ventrally-directed occipital
condyle are all features found in Anodontosaurus lambei and Euoplocephalus tutus (Arbour & Currie
2013a; Vickaryous & Russell 2003), Saichania chulsanensis (Maryańska 1977), and Tsagantegia
longicranialis (Tumanova 1993), for example. The lack of fusion between the quadrate and the

paroccipital process, and the orientation of the pterygoid body, differentiate INBR1004 from Saichania
chulsanensis (Maryańska 1977), but not from Ankylosaurus magniventris (Carpenter 2004) or
Euoplocephalus tutus (Arbour & Currie 2013a). Foramina for premaxillary and maxillary sinuses in the
external nares are present in Pinacosaurus grangeri (Maryańska 1977), P. mephistocephalus (Godefroit
et al. 1999), and Saichania chulsanensis (Maryańska 1977). The premaxilla forms the anterior rim of the
palatal vacuity in Pinacosaurus grangeri (Maryańska 1977) and INBR21004 (Miles & Miles 2009).
The referral of INBR21004 (holotype of Minotaurasaurus ramachandrani), but not PIN
3142/250, to Tarchia kielanae dramatically changes our understanding of the genus Tarchia and its
diagnosis. The referred skull INBR21004 provides much of the information used to differentiate Tarchia
kielanae from Saichania chulsanensis, also known from the Baruungoyot Formation. In addition to the
revised diagnostic characters proposed for Tarchia kielanae, this taxon can be differentiated from
Saichania chulsanensis by the presence of numerous small lacrimal and loreal caputegulae (only one pair
of lacrimal and one pair of loreal caputegulae in Saichania chulsanensis), a larger and more pyramidal
prefrontal caputegulum, anterior and posterior supraorbital caputegulae with distinct peaks and
separated by a notch in dorsal view (continuous edge in Saichania chulsanensis), less prominent nuchal
caputegulae, a mandibular caputegulam that extends nearly the entire length of the mandible
(caputegulum only extends about 75% of the mandible in Saichania chulsanensis), and a visible occipital
condyle in dorsal view. Tarchia kielanae has a constriction in the snout anterior to the orbits ('lacrimal
incisure' sensu Hill, Witmer & Norell, 2003), which is absent in Saichania chulsanensis but present in
Pinacosaurus grangeri (Fig. 5) Tarchia kielanae has narrower, more conical squamosal horns compared
to Saichania chulsanensis, but similar to those of Pinacosaurus mephistocephalus; the squamosal horns
of Pinacosaurus mephistocephalus are smoother than those of Tarchia kielanae, but it is unclear if this
could be related to ontogenetic stage. Although numerous isolated postcranial specimens are known
from the Baruungoyot Formation, at present none can be referred to Tarchia kielanae because of the
lack of overlapping cranial material, and because Tarchia kielanae is not the only ankylosaurid known
from this formation. Recently, a small skull with some associated postcrania from the Djadokhta
Formation has been referred to Minotaurasaurus ramachandrani (Alicea & Loewen, 2013).
SAICHANIA CHULSANENSIS MARYAŃSKA, 1977
Holotype: MPC 100/151, complete skull and both mandibles, seven cervical vertebrae (including fused
atlas and axis), ten dorsal vertebrae, ribs, sternum, both scapulocoracoids, humerus, ulna, radius,
manus, osteoderms including first and second cervical half rings; cast of specimen before individual
elements were separated at ZPAL.
Holotype locality and stratigraphy: Khulsan, Mongolia; Baruungoyot Formation (Mid-Upper Campanian,
Jerzykiewicz, 2000)
Referred specimens: PIN 3142/250, complete skull, both mandibles, and predentary (described by
Tumanova 1977), plus undescribed cervical vertebrae, scapula, sacrum, ischia, femur, ribs, and
osteoderms (some osteoderms on display at PIN), from Hermiin Tsav I, Mongolia, Nemegt Formation,
(Upper Campanian – Lower Maastrichtian, Jerzykiewicz, 2000). PIN 3142/251, complete skeleton with
skull, referred to Saichania chulsanensis by Tumanova (1987), tail club on display at PIN but rest of
skeleton undescribed and unfigured ( from Hermiin Tsav II); ZPAL MgD I/114, fragment of skull roof and

osteoderms, referred to Saichania chulsanensis by Maryańska (1977) but not described or figured and
could not be located during a visit in 2009 to ZPAL (from Hermiin Tsav II).
Revised Diagnosis: Ankylosaurine ankylosaurid with bulbous cranial ornamentation. Uniquely among
ankylosaurines, has one flat internarial caputegulum; one opening for cranial nerves IX-XII; fused atlas
and axis forming a syncervical; proximally wide humerus (proximal width 70% total humerus length);
intercostal ossifications present (may also be present in MPC 100/1305); and cervical half rings
composed of the underlying band, primary osteoderms, and coossified interstitial osteoderms
completely obscuring the band in external view.
Previous diagnoses: From Maryańska (1977): Large, oval external nostrils situated terminally, divided by
horizontal septum; premaxillary sinus present; premaxillary portion of rostrum relatively narrow;
premaxillae partly covered by well-developed ornamentation; occipital condyle weakly convex, ventrally
directed; epipterygoid present; exoccipital low, perpendicular to skull roof, ventral part deflected
anteriorly; quadrate oblique with condyle at level of middle part of orbit; orbits anteriorly and
posteriorly closed by partly neomorphic bones; skull roof overhangs occipital region; palatal region with
strongly developed anterior and posterior maxillary shelves; main body of maxilla surrounds palatal
vacuities over small area laterally; one opening for nerves IX-XII; atlas and axis fused; strongly developed
intercostal ossifications on trunk; limb bones very massive; forelimb strongly flexed; manus pentadactyl.
From Carpenter et al. (2011): cranial ornamentation of large protuberances; squamosal horn large and
triangular, contacts quadratojugal horn in occipital view; ridge-like, overhanging supraorbitals; external
nares laterally flaring; deeply recessed nasal vestibule with multiple sinus foramina; orbit located at mid-
length of skull; paroccipital process L-shaped in vertical cross-section; tooth rows divergent posteriorly
only; cervical neural arches X-shaped in dorsal view with low neural spines; dorsal centra long and low
with pleurofossa; cervical armour with larger, posteriorly projecting, triangular osteoderms.
Status: Valid.
Discussion: The holotype specimen of Saichania chulsanensis, MPC 100/151, includes the articulated
front half of the animal with skull and in situ osteoderms. A cast of the articulated specimen was made
prior to final preparation. The skull was described in detail by Maryańska (1977) and Carpenter et al.
(2011). Several of the diagnostic features noted by Maryańska (1977) have a broader distribution within
Ankylosauridae: large, oval external nares; a convex, ventrally directed occipital condyle; low
paroccipital processes oriented perpendicular to the skull roof; the orientation of the quadrate; and
strongly developed maxillary shelves are present in Ankylosaurus magniventris (Carpenter 2004),
Anodontosaurus lambei and Euoplocephalus tutus (Arbour & Currie 2013a; Vickaryous & Russell 2003),
Pinacosaurus grangeri (Maryańska 1977), and Tsagantegia longicranialis (Tumanova 1993). As in
Saichania chulsanensis, the orbits of Euoplocephalus tutus are closed anteriorly and posteriorly by bony
walls (Arbour & Currie 2013a). A premaxillary sinus, and premaxillae partly covered by well-developed
ornamentation, are also present in INBR21004 (Minotaurasaurus ramachandrani; Miles & Miles 2009).
In addition to Saichania chulsanensis, epipterygoids are known in Euoplocephalus tutus and
Pinacosaurus grangeri (Vickaryous et al. 2004). Massive limb bones, a strongly flexed forelimb, and
pentadactyl manus are also present in Pinacosaurus grangeri (Maryańska 1977; Maleev 1954) and
Anodontosaurus lambei (Coombs 1986). The premaxillary portion of the rostrum is not noticeably
narrower compared to INBR21004 (Minotaurasaurus ramachandrani, Miles & Miles 2009) or
Anodontosaurus lambei and Euoplocephalus tutus (Arbour & Currie 2013a; Vickaryous & Russell 2003).

Carpenter et al. (2011) revised the diagnosis for Saichania chulsanensis, including some new
diagnostic characters, but many of these can also be found in other ankylosaurids. Large triangular
squamosal horns and ridge-like, overhanging supraorbitals are present in Ankylosaurus magniventris
(Carpenter 2004), Euoplocephalus tutus, and Scolosaurus cutleri (Arbour & Currie 2013a; Vickaryous &
Russell 2003). Bulbous cranial ornamentation and flaring external nares are also present in
Minotaurasaurus ramachandrani (Miles & Miles 2009). A deep nasal vestibule with multiple foramina is
present in Minotaurasaurus ramachandrani (Miles & Miles 2009) and Pinacosaurus grangeri (Maryańska
1977; Hill et al. 2004). Cervical osteoderms with large triangular osteoderms are present in
Euoplocephalus tutus and Scolosaurus cutleri (Arbour & Currie 2013a), but the cervical osteoderms of
MPC 100/151 do appear to posteriorly overhang the underlying bony band somewhat more than in
other ankylosaurs. Other characters proposed by Carpenter et al. (2011) are arguable based on
anatomical interpretations. Carpenter et al. (2011) noted that the orbit was positioned more anteriorly
in the skull of MPC 100/151 compared to many other ankylosaurids. In lateral view, the orbit of MPC
100/151 is more anteriorly placed compared to that of PIN 3142/250 (Fig. 5). However, in
Anodontosaurus lambei and Euoplocephalus tutus, the relative position of the orbit seems to be
influenced in part by the length of the rostrum, which may result from taphonomic deformation in the
dorsoventral plane 'lengthening' the snout as the arched rostrum is flattened (Arbour & Currie 2013a:
Fig. 5). A contact between the squamosal and quadratojugal horns in posterior view was also proposed
as a diagnostic feature, but examination of the skull of MPC 100/151 shows that the squamosal and
quadratojugal horns do not contact each other, and are instead separated by a gap and a single
postocular osteoderm. Similarly, the tooth rows of MPC 100/151 are not only divergent posteriorly, but
anteriorly as well, although the divergence is not as great anteriorly as posteriorly. The remaining two
postcranial characters were discussed by Arbour & Currie (2013b): X-shaped cervical neural arches are
also found in Euoplocephalus tutus; the cervical neural spines are broken, not low, in MPC 100/151;
long, low dorsal centra are present in many ankylosaurids; and the shallow depressions on the lateral
sides of the dorsal centra most likely do not represent pleurofossae, a feature otherwise known only in
saurischian dinosaurs.
Saichania chulsanensis differs significantly from Tarchia kielanae, particularly with respect to the
cranial ornamentation, but shares (Figs. 4, 5) several features with PIN 3142/250, which was previously
referred to Tarchia gigantea. Tumanova (1977) reassigned Dyoplosaurus giganteus to Tarchia (forming
Tarchia gigantea) because the holotypes of both Dyoplosaurus giganteus and Tarchia kielanae were
from coeval deposits from geographically close localities, and because of similarities in osteoderm
shapes between PIN 551/29 (holotype of Dyoplosaurus giganteus) and the newly-collected PIN
3142/150 (although the osteoderms of PIN 3142/150 were not figured). Most discussions of the
morphology and systematics of Tarchia gigantea focus on PIN 3142/250, and most character codings in
phylogenetic analyses rely heavily on this specimen (Vickaryous et al., 2004; Thompson et al., 2011). The
holotype of Dyoplosaurus giganteus lacks autapomorphies or a unique combination of characters and is
a nomen dubium, and PIN 3124/250 differs in several respects from ZPAL MgD I/111 (Fig. 3). As such,
PIN 3124/250 cannot be referred to Tarchia kielanae or Tarchia gigantea. Instead, PIN 3142/250 is here
referred to the genus Saichania chulsanensis.
MPC 100/151 (Saichania chulsanensis) and PIN 3142/250 both have pyramidal squamosal horns,
prominent nuchal caputegulae, a generally similar pattern of bulbous frontonasal caputegulae, large,

flat, rectangular lacrimal caputegulae and large, flat, rectangular loreal caputegulae (Figs. 4, 5). The two
skulls differ in some aspects of the cranial ornamentation. The internarial caputegulum is anteriorly
forked and more rugose in PIN 3142/250 than in MPC 100/151, and the rugose premaxillary
ornamentation is limited to the most anterior parts of the premaxillae at the midline of the skull in PIN
3142/250. The supranarial caputegulae are anteroposteriorly narrower in PIN 3142/250 than in MPC
100/151. The skulls also differ in the relative sizes of the prefrontal caputegulum (smaller in MPC
100/151), middle supraorbital caputegulum (larger and rounder in MPC 100/151) and squamosal horn
(deeper in MPC 100/151), and in the presence or absence of postocular caputegulae (present in MPC
100/151, absent in PIN 3142/250). The nuchal caputegulae are less prominent in PIN 3142/250 than in
MPC 100/151. The quadrate and paroccipital process are coossified in MPC 100/151, but are not in PIN
3142. Finally, the rostrum of PIN 3142/250 is relatively longer than in MPC 100/151; the orbit is situated
at about the midlength of the skull in MPC 100/151, but is more posterior in PIN 3142/250.
Although several differences between PIN 3124/250 and the holotype of Saichania chulsanensis
(MPC 100/151) have been noted previously (Tumanova, 1987; Carpenter et al., 2011), some of these
differences may result from intraspecific variation. This is difficult to assess with only two specimens, as
any particular variation will, by default, represent the ends of a potential continuum of variation. Only a
few ankylosaurids are known from enough specimens to assess intraspecific or ontogenetic variation:
the North American ankylosaurids Anodontosaurus lambei, Euoplocephalus tutus, and Scolosaurus
cutleri, and the Asian Pinacosaurus grangeri. A review of variation within Anodontosaurus lambei,
Euoplocephalus tutus and Scolosaurus cutleri found that the overall pattern of frontonasal
ornamentation, the depth of the squamosal horn, and the presence or absence of postocular
caputegulae did not vary intraspecifically (Arbour & Currie, 2013a). The exact shapes, sizes, and
placements of individual frontonasal and frontoparietal caputegulae varied greatly among specimens, as
did the prominence of the supranarial caputegulae. Squamosal horn length and bluntness varied within
Anodontosaurus lambei, Euoplocephalus tutus, and Scolosaurus cutleri, although Scolosaurus cutleri had
longer and sharper squamosal horns relative to Anodontosaurus lambei and Euoplocephalus tutus.
Taxonomically significant features of the skull included the presence or absence of postocular
caputegulae, and the shapes of the squamosal horns (Arbour & Currie, 2013a).
Based on variation within the North American taxa, the variation in the anteroposterior length
of the supranarial caputegulae, and the size of the middle supraorbital caputegulae in PIN 3142/250 and
MPC 100/151, may represent intraspecific variation. The presence or absence of postocular caputegulae
was taxonomically significant for Anodontosaurus lambei and Euoplocephalus tutus, and so the presence
of these caputegulae in MPC 100/151 and their absence in PIN 3142/250 may be important. Although
squamosal horn length can vary within Euoplocephalus tutus, squamosal horn height at the base does
not seem to vary greatly (Arbour & Currie, 2013a), and so the difference in squamosal horn depths in
MPC 100/151 and PIN 3142/250 may also be taxonomically significant.
In addition to variation in the cranial ornamentation, there are several differences in the
morphology of the skull of MPC 100/151 and PIN 3142/250. In MPC 100/151, the quadrates are
coossified to the paroccipital processes, whereas the quadrate and paroccipital process are unfused in
PIN 3140/250. This has previously been interpreted as a taxonomically significant feature (Tumanova,
1987). However, fusion of skeletal elements can be related to ontogenetic state. Although most
specimens of Anodontosaurus lambei and Euoplocephalus tutus do not show coossification of the

quadrate and paroccipital process, at least one does (TMP 1997.132.1; see Arbour & Currie, 2013a: Fig.
6).The basipterygoid processes in PIN 3142/250 are anteroposteriorly short, but distinct, nubs. In
contrast, the basipterygoid processes in MPC 100/151 are not distinctly separated, and this region is
rugose. The morphologies of the basipterygoid processes in Anodontosaurus lambei and Euoplocephalus
tutus are similar among all referred specimens (Arbour & Currie, 2013a), and so the difference in
morphology between PIN 3142/250 and MPC 100/151 may be taxonomically significant.
Overall, MPC 100/151 and PIN 3142/250 share a large number of morphological features in
common, and most differences between these specimens probably result from individual variation.
These specimens represent at least the same ankylosaurid genus. However, a few differences between
MPC 100/151 and PIN 3142/250 fall outside the range of variation observed in other better represented
ankylosaurids: the presence or absence of postocular caputegulae, the depth of the squamosal horns,
and the morphology of the basipterygoid processes. Coossification of the quadrate and paroccipital
process may also be taxonomically significant. Until additional specimens of potential Saichania
chulsanensis skulls are described that may clarify some of these potential taxonomic differences, PIN
3142/250 will be referred to Saichania chulsanensis.
PIN 3142/250 has an unusual hole dorsal to the right orbit; remodeled bone along its rim
indicates that this is pathological, not taphonomic, in nature (Fig. 4). The hole pierces directly into the
orbital cavity. Gallagher et al. (1998) observed a zoned bony growth in the nasal passages between the
internal nares and right orbit using computed tomography scans. They suggested that the bony mass
and hole may represent a healing puncture wound, possibly from a tyrannosaur bite. Further
assessment of this unusual feature of PIN 3142/250 awaits a full description of the computed
tomography scan results.
Saichania chulsanensis has some unique postcranial features compared to other ankylosaurids.
MPC 100/151 is the only ankylosaurid known to possess a fused atlas-axis complex. MPC 100/151 also
has a particularly robust humerus, with a proximal width 70% of the humeral length; Arbour & Currie
(2013b) suggest that this was unlikely to be size-related. The cervical half rings are the most elaborate of
any known ankylosaurid. In addition to the typical six, keeled major osteoderms, the band is almost
completely covered by coossified interstitial osteoderms (Arbour & Currie, 2013b). There is a
proportionately larger conical osteoderm between the lateral and distal osteoderms of the second
cervical half ring, which forms the centre of a rosette. Although some other ankylosaurids have
interstitial osteoderms on the cervical half rings (Anodontosaurus lambei, Scolosaurus cutleri), these are
usually found mostly at the bases of the major osteoderms, and do not form extensive coossified sheets
above the band (Arbour & Currie, 2013a).
Several specimens have been referred to Saichania chulsanensis, including MPC 100/1305, a
nearly complete postcranial skeleton lacking a skull collected by a Russian-Mongolian expedition in 1976
(Arbour & Currie 2013b). The skull described for this specimen by Carpenter et al. (2011) is a cast of MPC
100/151; Arbour & Currie (2013b) argued that differences in aspects of the postcranial morphology
between MPC 100/151 and MPC 100/1305 indicate that MPC 100/1305 should not be referred to
Saichania chulsanensis. Additionally, MPC 100/1305 was found to have been collected from the
Djadokhta Formation rather than the originally reported Baruungoyot Formation. At present, only
Pinacosaurus grangeri and Pinacosaurus mephistocephalus have been identified from the Djadokhta
Formation, as well as a possible specimen of Tarchia kielanae (as Minotaurasaurus ramachandrani,

Alicea & Loewen 2013). ZPAL MgD I/114, a fragment of skull roof and osteoderms, was referred to
Saichania chulsanensis by Maryańska (1977) but could not be located at the ZPAL collections during a
visit by VMA in 2009. Tumanova (1987) referred PIN 3142/251, a nearly complete skeleton with skull, to
Saichania chulsanensis, but this specimen is undescribed (although the tail club is on display at the PIN).
These specimens may provide additional information on the anatomy of Saichania chulsanensis.
The referral of PIN 3142/250 to Saichania chulsanensis extends the stratigraphic range of
Saichania chulsanensis. At present, postcranial specimens from the Nemegt Formation previously
referred to Dyoplosaurus giganteus cannot be referred to Saichania chulsanensis. This is unfortunate,
because at least two tail club morphotypes are represented in the Nemegt Formation. PIN 551-29
(holotype of Dyoplosaurus giganteus), ZPAL MgD I/42 (free caudal vertebrae and tail club), and ZPAL
MgD I/43 (the largest known tail club knob for any ankylosaur) have handle caudal vertebrae with V-
shaped neural spines, in which the prezygapophyses diverge at an angle of about 22-26° (Fig. 2E), the
typical condition for most ankylosaurines (Arbour et al., 2009). ZPAL MgD I/113 has handle vertebrae in
which the prezygapophyses diverge at an angle of about 37°, an intermediate morphology between the
typical V-shaped morphology of taxa like Euoplocephalus tutus or Pinacosaurus grangeri, and the U-
shaped morphology of Ankylosaurus magniventris (Arbour et al., 2009; Fig. 2F). This unusual morphology
is unlikely to represent intraspecific variation, and so ZPAL MgD I/113 probably does not represent the
same taxon as PIN 551-9, ZPAL MgD I/42, or ZPAL MgD I/43 (Arbour et al., 2013). Once the postcranium
of PIN 3142/250 is described, it may be possible to refer either the V-shaped or intermediate
morphology handle vertebrae to Saichania chulsanensis, but at present, the lack of overlapping material
prevents assignment of either tail club morphotype to Saichania chulsanensis. However, tail clubs
recovered from the Nemegt Formation indicate that at least two ankylosaurine species were present in
this formation.
ZARAAPELTA NOMADIS GEN. ET SP. NOV.
Holotype: MPC D100/1338, a partial skull missing the rostrum.
Etymology: Zaraapelta nomadis, зараа (Mongolian) hedgehog, in reference to the spiky appearance of
the skull, and pelta (Latin), a small shield, in reference to the osteoderms found on all ankylosaurs;
nomadis, from nomas (Latin), nomad, in reference to Mongolian travel company Nomadic Expeditions,
which has facilitated many years of palaeontological fieldwork in the Gobi Desert.
Holotype Locality and Horizon: N43°28.345’, E99°51.032’ (WGS 84), Hermiin Tsav, Gobi Desert,
Mongolia; Baruungoyot Formation (Mid-Upper Campanian, Jerzykiewicz, 2000).
Diagnosis: Ankylosaurine ankylosaurid with bulbous cranial ornamentation. Unlike other ankylosaurs,
squamosal horn has unique smooth-textured keel offset from the rest of the squamosal horn by a
distinct and abrupt change to a granular texture; elaborate pattern of postocular caputegulae covering
entire postocular region between squamosal and quadratojugal horns, with more postocular
caputegulae than Anodontosaurus lambei, Saichania chulsanensis or Tarchia kielanae.
Description: MPC D100/1338 (Figs. 4-5, 8-11) is a partial skull missing the portion of the rostrum anterior
to the prefrontals. The left side is otherwise complete, but the right side of the skull is broken across the
orbit and lacks the pterygoid, quadrate, quadratojugal, and jugal. No teeth are preserved in situ. The
antorbital fenestra is absent and the laterotemporal fenestra is obscured in lateral view by the
squamosal and quadratojugal. The skull bears the characteristic ankylosaurid cranial sculpturing on the

dorsal surface, including prominent squamosal and quadratojugal horns. The description of this skull
follows the regional terminology proposed by Vickaryous & Russell (2003) wherein the skull is
subdivided into rostral, temporal, palatal, and occipital/basicranial regions.
Rostral Region – Both maxillae are badly damaged and missing their anterior ends. The tooth
row is inset from the lateral side of the maxilla, and extends posteriorly almost to the pterygoid. At the
posterior end of the maxilla, dorsal to the contact with the pterygoid flange, is a posteriorly-directed
circular aperture for the maxillary artery (Fig. 10). The nasals are poorly preserved, and the posterior
extents of the nasals are unknown. Ventrally, the nasals contribute to the median nasal septum.
Anteriorly, the broken edge of the rostrum reveals three openings in each nasal. These openings lead to
channels within the nasal that are confluent with a more posteriorly-placed, posteriorly-oriented
opening on the ventral side of the skull roof (Fig. 11). These channels may represent parts of the nasal
passages, which in Euoplocephalus tutus are complex and looping (Witmer & Ridgely, 2008).
The prefrontal caputegulum in Zaraapelta nomadis is a large, prominent triangle in dorsal view
(Fig. 9), is slightly concave on its dorsal surface, and is most similar to that of Tarchia kielanae
(INBR21004). It differs from the prefrontal caputegulae in Saichania chulsanensis (MPC 100/151),which
are keeled and in dorsal view have straight edges. In Pinacosaurus grangeri (AMNH 6523) the prefrontal
caputegulum is prominent and sharply pointed. The boundaries of the lacrimal are not visible in MPC
D100/1338, but in Pinacosaurus grangeri (ZPAL MgD II/1) this bone forms the anterior edge of the orbit.
The frontals are obscured by the frontonasal caputegulae, which are bulbous and pyramidal. In
dorsal view, two pairs of transversely-oriented, rectangular to trapezoidal frontonasal caputegulae are
present on each side of the midline of the skull (Fig. 8). Lateral to these are smaller, roughly square
frontonasal caputegulae. The posterior region of the frontals does not have discrete ornamentation, but
discrete frontonasal caputegulae are preserved anterior to the middle supraorbitals. Discrete
caputegulae are present between the middle supraorbitals in Saichania chulsanensis (MPC 100/151) and
Tarchia kielanae (INBR21004), but are not present in this region in Zaraapelta nomadis (Fig. 4); instead,
the most posterior discrete caputegulae in Zaraapelta nomadis are present anterior to the anterior edge
of the middle supraorbital. Although the development of cranial ornamentation may proceed anteriorly
to posteriorly (compare the Pinacosaurus grangeri juvenile specimen ZPAL MgD II/1 with the adult
specimen AMNH 6523, Fig. 4), the absence of caputegulae in the posterior frontal region of Zaraapelta
nomadis is probably not ontogenetically related, as the holotype of Zaraapelta nomadis is considerably
larger than that of INBR21004 (Tarchia kielanae). Ventrally, the frontal has a scroll-like, descending
process (Figs. 10, 11), similar to that observed in some specimens of Euoplocephalus tutus (AMNH 5238,
UALVP 47977); this may represent the posterior wall of the olfactory turbinate (Miyashita et al., 2011).
Unlike in Euoplocephalus tutus, there is no groove associated with the descending process in the nasal
cavity (Miyashita et al., 2011). In contrast, ZPAL MgD I/111 (Tarchia kielanae) does not appear to have a
scroll-like, descending process on the frontal (Maryańska, 1977: Pl. 24, Fig. 2).
Temporal Region – Dorsal to the orbit are three supraorbital caputegulae: a smaller anterior
supraorbital, a larger posterior supraorbital, and a middle supraorbital positioned more medially that
does not reach the lateral edge of the skull (Fig. 9). The posterior supraorbital caputegulum, in dorsal
view, is more triangular than the anterior supraorbital caputegulum, but the lateral edge is concave. In
dorsal view, the lateral edges of the anterior and posterior supraorbitals form a continuous edge in
Saichania chulsanensis (MPC 100/151); Zaraapelta nomadis is more similar to Pinacosaurus grangeri

(but not Pinacosaurus mephistocephalus, Godefroit et al. 1999) and Tarchia kielanae in the
morphologies of the supraorbitals, in which each supraorbital has a distinct apex in dorsal view (Fig. 4).
No eyelid ossifications (known in Dyoplosaurus acutosquameus and Euoplocephalus tutus; Coombs
1972; Arbour & Currie 2013a) were preserved in MPC D100/1338. The boundaries of the postorbital
cannot be distinguished, but in Pinacosaurus grangeri the postorbital contributes to the postocular
shelf. The jugal forms the ventral border of the orbit; as in all ankylosaurs it is shallow, but it is
proportionately deeper in Zaraapelta nomadis than the jugals in Pinacosaurus grangeri (IVPP V16853,
ZPAL MgD II/1), Pinacosaurus mephistocephalus (Godefroit et al. 1999), , and Tarchia kielanae (ZPAL
MgD I/111; Fig. 5).
The parietals in MPC D100/1338 lack distinct caputegulae and are smooth with sparse pitting
(Fig. 9). Posteriorly, the parietals form a nuchal shelf that nearly obscures the braincase in dorsal view; a
small crescent of the supraoccipital is visible in dorsal view. This is similar to the condition in Tarchia
kielanae (ZPAL MgD I/111) but differs from Pinacosaurus grangeri (ZPAL MgD II/1) and Saichania
chulsanensis (MPC 100/151), in each of which the braincase is completely obscured by the nuchal shelf
in dorsal view (Fig. 4). The nuchal shelf is fused to the supraoccipital and paroccipital processes. In
Saichania chulsanensis (MPC 100/151) and Tarchia kielanae (ZPAL MgD I/111), the nuchal shelf has two
distinct caputegulae. In MPC D100/1338, there are two caputegulae, but the anterior border is weakly
developed relative to those in Saichania chulsanensis (MPC 100/151) and Tarchia kielanae (ZPAL MgD
I/111). In Euoplocephalus tutus, the nuchal shelf has four to six discrete caputegulae (Arbour & Currie
2013a).
The squamosal forms the dorsal posterolateral corner of the skull, and is developed into the
characteristically ankylosaurid pyramidal squamosal horn. The squamosal horn of MPC D100/1338 is
unique among ankylosaurs. The dorsal keel is sharp and the immediately surrounding bone texture is
smooth. A short distance from the keel, there is a distinct edge that demarcates a change in texture
from smooth to granular (Figs. 8, 9). The squamosal horns of Saichania chulsanensis (MPC 100/151, PIN
3142/250) have uniform textures (Fig. 5). The squamosal horn in Zaraapelta nomadis is pyramidal, as in
Saichania chulsanensis, not slender as in Pinacosaurus mephistocephalus (Godefroit et al. 1999) or
Tarchia kielanae (INBR21004) (Fig. 4). The squamosal horns of Tarchia kielanae (INBR21004) are narrow
and project well beyond the posterior margin of the skull. Tarchia kielanae also uniquely has a set of
accessory, elongate caputegulae (or possibly osteoderms) anterior to the squamosal horns, on top of
the postorbitals. One potential explanation for the unusual bi-layered texture of Zaraapelta nomadis is
that the postorbital osteoderm and squamosal horn of Tarchia kielanae fuse together during ontogeny,
and that Zaraapelta nomadis represents a more mature specimen of Tarchia kielanae. INBR21004 still
has visible supraorbital-frontal and squamosal-parietal sutures on the dorsal surface of the skull, unlike
most other known ankylosaurid adults; this is similar to juvenile Pinacosaurus grangeri (Maryańska
1977; Burns et al. 2011) which suggests that this specimen does not represent a fully mature individual
(contra Carpenter et al. 2011). However, the squamosal horns of Pinacosaurus grangeri are
anteroposteriorly short and blunt, in both juveniles (ZPAL MgD II/1) and adults (AMNH 6523; Fig. 4). The
similarity of squamosal horn shape in juvenile and adult Pinacosaurus grangeri suggests that squamosal
horn shape does not change dramatically throughout ontogeny. Additionally, the differences between
the squamosal horns of Tarchia kielanae and Zaraapelta nomadis are well beyond the range of
morphological variation observed in the relatively large sample sizes of the North American genera

Anodontosaurus lambei, Euoplocephalus tutus, and Scolosaurus cutleri (Arbour & Currie, 2013a), which
supports separation of ZPAL MgD I/111 and MPC D100/1338 as distinct taxa.
The quadratojugal forms the ventral posterolateral corner of the skull, and the quadratojugal
horn is most likely an outgrowth of the quadratojugal (Vickaryous et al., 2001). In lateral view the
quadratojugal horn is triangular, with a slightly concave posterior border (Fig. 8). In Tarchia kielanae
(INBR21004), there is a distinct notch at the anterior and proximal edge of the quadratojugal horn, but
in MPC D100/1338 the quadratojugal horn contacts the jugal in this region (Fig. 5). The quadratojugal
horn has a smoother texture than the granular region of the squamosal horn, with a few shallow
grooves radiating from the apex of the horn. The quadratojugal horn obscures the quadrate in lateral
view.
Immediately posterior to the orbit are six caputegulae separated by shallow furrows; dorsally,
these are long and rectangular, but they decrease in length ventrally (Fig. 8). Posterior to this set of
caputegulae are seven smaller, square caputegulae separated by deep furrows, and which are more
irregularly arranged. A large and particularly prominent triangular caputegulum is present at the
posterior edge of the skull in lateral view. Three indistinct bumps are present at the very base of the
squamosal horn. Some other ankylosaurids also have caputegulae in this region of the skull, such as
Anodontosaurus lambei (Arbour & Currie 2013a), and Tarchia kielanae (INBR21004). However, in no
other ankylosaurid are these smaller caputegulae as abundant or prominent.
Palatal Region – The vomers are badly damaged, but may have extended nearly to the level of
the maxillary tooth rows (Fig. 10). The nasal passages were probably completely subdivided by the
nasals and vomers. The palatines are not preserved. The right pterygoid is not preserved and the left
pterygoid is damaged. The pterygoid body is transversely and vertically oriented. It is slightly concave
ventrally, and a pterygoid foramen is present. The pterygoid body is fused to the basipterygoid process,
as in Saichania chulsanensis (MPC 100/151). The pterygoid flange projects anterolaterally; it is not tilted
anteriorly as in Tarchia kielanae (INBR21004), but has a horizontally flat ventral surface. Although the
right pterygoid is not preserved, it appears that a substantial interpterygoid vacuity was present. The
quadrate ramus projects posterolaterally and is fused to the quadrate (although the edges of the scarf
joint are still slightly visible). The ectopterygoid cannot be distinguished.
Occipital/Basicranial Region – The supraoccipital is an unpaired median bone dorsal to the
foramen magnum. It is fused to ventral surface of the parietals, and has a pair of low dorsal
prominences (Fig. 11). The exoccipital and opisthotic are fused and form the paroccipital process. The
exoccipital contributes to the lateral wall of the foramen magnum, and the opisthotic contributes to the
lateral wall of the endocranial cavity. The exoccipitals contact the basioccipital ventromedially. The
paroccipital process extends laterally from the foramen magnum, and the lateral terminus is fused with
the quadrate. This is similar to the condition in one specimen of Saichania chulsanensis (MPC 100/151)
but not in Tarchia kielanae (INBR21004) where the paroccipital process and quadrate are not fused. The
paroccipital process does not fuse to the squamosal. In posterior view, the paroccipital process is
somewhat downturned with a concave ventral surface. In dorsal view, the paroccipital processes are
obscured by the nuchal crest. The basioccipital forms the posterior floor of the braincase. The occipital
condyle is formed only by the basioccipital, and is the typical ankylosaurid reniform shape in posterior
and ventral views. The occipital condyle is not offset from the rest of the basioccipital by a neck. The
posterior edge of the occipital condyle is visible in dorsal view in Zaraapelta nomadis, unlike the

conditions in almost all ankylosaurids, but similar to that of Tarchia kielanae (INBR21004). However, in
Tarchia kielanae (INBR21004) much more of the occiput is visible in dorsal view, including more of the
occipital condyle and the paroccipital processes.
Anterior to the basioccipital is the unpaired basisphenoid. The contact between the basioccipital
and basisphenoid is transversely through the basal tubera, which takes the form of a rugose transverse
ridge on the ventral surface of the braincase (Fig. 10). Anteriorly, the basisphenoid bifurcates into stout,
anterolaterally-directed basipterygoid processes. The basipterygoid process fuses to the posterior face
of the pterygoid body. There are no sutural contacts that mark the boundaries of the prootic and
opisthotic, but these bones form the lateral walls of the braincase.
The limits of the individual bones in the sphenoid region, as in many dinosaurs, are difficult to
discern because of extensive fusion. The laterosphenoid is fused to the skull roof and contributes to the
postocular shelf, which is weakly developed in MPC D100/1338 relative to other ankylosaurids like
Euoplocephalus tutus (Miyashita et al., 2011). It is dorsoventrally shallow and extends approximately
half the distance from the braincase to the lateral edge of the skull. The orbitosphenoid contributes to
the lateral wall of the braincase and the medial wall of the orbit. The parasphenoid forms the
anteroventral floor of the braincase and is indistinguishably fused to the basisphenoid. The
parasphenoid tapers anteriorly into the long, triangular parasphenoid rostrum (cultriform process),
which supports the interorbital septum. The mesethmoid and sphenethmoid are not visible in MPC
D100/1338. The ectethmoid is a thin sheet of bone that separates the orbit from the olfactory region. It
forms a horizontal shelf ventral to the scroll-like descending process of the frontal, and obscures the
descending process in ventral view.
The quadrate is posterior to the pterygoid and anterior to the paroccipital process. It is fused to
the paroccipital process, quadratojugal, and pterygoid. The quadratojugal overlaps the ventrolateral
edge of the quadrate dorsal to the articular condyle. In ventral view, the articular condyle of the
quadrate is roughly triangular, and is wider medially than laterally. In posterior view, the ventral surface
of the articular condyle is weakly saddle-shaped.
ANKYLOSAURIDAE GEN. ET SP. INDET., FROM THE NEMEGT FORMATION OF MONGOLIA:
PIN 551/29, caudal vertebrae, metatarsals, phalanges, osteoderms including partial tail club knob,
holotype of Dyoplosaurus giganteus (Nemegt); PIN 5011/87, first cervical half ring, on display at PIN as
Tarchia; MPC KID 233 – undescribed ankylosaurid skeleton (Hermiin Tsav I); MPC KID 329 – undescribed
juvenile limb elements (Hermiin Tsav I); MPC KID 335 – caudal vertebra (Hermiin Tsav I); MPC KID 336 –
undescribed juvenile material (Hermiin Tsav I); MPC KID 373 – partial dentary (Hermiin Tsav I); MPC KID
399 – undescribed partial skeleton (partially excavated originally by Russian expedition in 1972; Hermiin
Tsav I); MPC KID 515 – dorsal vertebrae, pedal phalanx, and osteoderms (Altan Uul II); MPC KID 538 –
partial tail club handle (Altan Uul II); KID 586 – humerus (southwest of Bugeen Tsav); MPC KID 589 –
cervical half ring fragment (Khuree Tsav); MPC KID 591 – free caudal vertebra and osteoderms
(southwest of Bugeen Tsav); MPC KID 630 – humerus (southwest of Bugeen Tsav); MPC KID 636 – free
caudals, handle caudal, osteoderms (southwest of Bugeen Tsav); MPC KID 637 – free caudal, osteoderms
(southwest of Bugeen Tsav); numerous isolated osteoderms or clusters of osteoderms from MPC KID
expeditions; ZPAL MgD I/42, tail club (Altan Uul IV); ZPAL MgD I/43, tail club, housed at MPC (Altan Uul
IV); ZPAL MgD I/49, right humerus (Altan Uul IV); ZPAL MgD I/113, partial pelvis, nearly complete caudal

series including free caudal vertebrae and handle but missing tail club knob, osteoderms, skin
impressions (Altan Uul III).
DISCUSSION
The phylogenetic analysis produced 46 most parsimonious trees (Fig. 12). The strict consensus
tree has a consistency index (CI) of 0.577, and a retention index (RI) of 0.642. The strict consensus tree
has significantly better resolution of ankylosaurid interrelationships, better bootstrap support, and
better Bremer support compared to the analyses in Arbour & Currie (2013a). The strict consensus tree
shows a close relationship between Tarchia kielanae and Zaraapelta nomadis. In 91% of the most
parsimonious trees, Saichania chulsanensis is the sister taxon to Tarchia kielanae and Zaraapelta
nomadis. These taxa all share pyramidal frontonasal caputegulae with sharp edges, and all except one
specimen of Saichania chulsanensis (PIN 3142/250) have postocular caputegulae (also present in the
North American taxa Anodontosaurus lambei and Scolosaurus cutleri). Tarchia kielanae and Zaraapelta
nomadis share supraorbital caputegulae with distinct apices (also present in Pinacosaurus grangeri), a
pyramidal prefrontal caputegulum (also present in Pinacosaurus grangeri), and more than one
caputegulum in the lacrimal and loreal regions. Saichania chulsanensis, Tarchia kielanae, and Zaraapelta
nomadis form a polytomy with a clade of primarily North American ankylosaurids (Ankylosaurus
magniventris, Anodontosaurus lambei, Euoplocephalus tutus, and the Mongolian Talarurus
plicatospineus; recovered in 73% of all trees), Dyoplosaurus acutosquameus, Nodocephalosaurus
kirtlandensis, and Tianzhenosaurus youngi. This polytomy is the sister group to a monophyletic
Pinacosaurus (recovered in 91% of all trees). Scolosaurus cutleri, Tsagantegia longicranialis, and
Gastonia burgei formed successive sister taxa to this larger clade. Safe taxonomic reduction was
performed with TAXEQ3, but no species could be safely removed without removing useful phylogenetic
data. A maximum agreement subtree was calculated in TNT, and Dyoplosaurus acutosquameus,
Nodocephalosaurus kirtlandensis, and Tianzhenosaurus youngi were pruned from the resulting tree.
The results of this paper provide strong support for a Baruungoyot Formation origin for
INBR21004 (holotype of Minotaurasaurus ramachandrani, now referred to Tarchia kielanae). Miles &
Miles (2009) state that the matrix around the specimen suggested an origin in the Gobi Desert of
Mongolia or China. The referral of INBR21004 to Tarchia kielanae, the holotype of which was collected
from the Baruungoyot Formation at Khulsan, strongly suggests that this specimen derives from the
Baruungoyot Formation of the Mongolian Gobi. Additionally, the close relationship between Tarchia
kielanae and Zaraapelta nomadis (also from the Baruungoyot Formation of the Mongolian Gobi) lends
further support to the possible provenance of INBR21004 from Mongolia. However, the report of a new
skull referable to "Minotaurasaurus" by Alicea & Loewen (2013) from the Djadokhta Formation suggests
that Tarchia kielanae may be present in both the Baruungoyot and Djadokhta formations; additional
commentary on the stratigraphic range of this taxon must await a full description of the new specimen.
Three ankylosaurids have been identified from the Baruungoyot Formation (Saichania
chulsanensis, Tarchia kielanae, and Zaraapelta nomadis), and at least two may have been present in the
Nemegt Formation based on tail club morphotypes (one of which may belong to Saichania
chulsanensis). Although the holotype of Saichania chulsanensis preserves some of the postcranium,
Tarchia kielanae and Zaraapelta nomadis are at present only known from cranial material. As such, it is

not possible to assign isolated ankylosaurid postcranial remains from the Baruungoyot Formation to any
of the named species.
Many formations worldwide include more than one ankylosaur species, but only a few
formations include three or more. The Aptian-Albian Cedar Mountain Formation includes at least six
ankylosaur species, although these are distributed through different members of the formation
(Carpenter et al., 2008). The Campanian Dinosaur Park Formation includes the nodosaurids Edmontonia
rugosidens Gilmore, 1930 and Panoplosaurus mirus (Ryan & Evans, 2005) and the ankylosaurids
Dyoplosaurus acutosquameus, Euoplocephalus tutus, and Scolosaurus cutleri, although Scolosaurus
cutleri from Alberta may be derived from the Oldman Formation instead (Arbour & Currie, 2013a). The
Baruungoyot Formation includes the ankylosaurids Saichania chulsanensis, Tarchia kielanae, and
Zaraapelta nomadis. If the single specimen of Scolosaurus cutleri from the Dinosaur Park Formation is
instead from the underlying Oldman Formation, then the Baruungoyot Formation preserves the greatest
diversity of ankylosaurid ankylosaurs in the world. In the Dinosaur Park Formation, the beak shapes of
nodosaurid ankylosaurs and ankylosaurid ankylosaurs may have reflected niche partitioning of food
resources (Mallon & Anderson, 2013). The beak morphology of Zaraapelta nomadis is unknown, but the
beak morphologies of Saichania chulsanensis and Tarchia kielanae did not differ substantially, and
dietary niche partitioning between ankylosaurids cannot alone explain the high diversity of
ankylosaurids in the Baruungoyot Formation. Although an indeterminate sauropod is known from the
Baruungoyot Formation (Weishampel et al., 2004), sauropods appear to have been rare components of
the Baruungoyot ecosystem, and because of their ability to access forage at heights unavailable to
ankylosaurs, it seems unlikely that sauropods and ankylosaurids would be in direct competition for food
resources. Ankylosaurids can be considered the dominant megaherbivore in the Baruungoyot
Formation, which was otherwise composed primarily of small herbivorous or omnivorous dinosaurs like
the ceratopsian Bagaceratops rozhdestvenskyi Maryańska & Osmolska, 1975, the pachycephalosaurid
Tylocephale gilmorei Maryańska & Osmolska, 1974, oviraptorids, alvarezsaurids, and avimimids, a small
carnivorous velociraptorine (Weishampel et al. 2004; Longrich et al. 2010), and small birds, lizards, and
mammals.
Alternately, ankylosaurid diversity in the Baruungoyot Formation may have been accomplished
through sexual selection. Saichania chulsanensis, Tarchia kielanae, and Zaraapelta nomadis have some
of the most elaborate cranial ornamentation of any ankylosaurids, with relatively more pronounced
frontonasal, squamosal, and quadratojugal ornamentation compared to earlier ankylosaurids from
Mongolia and China (e.g. Pinacosaurus grangeri, Tsagantegia longicranialis Tumanova, 1993) or
contemporaneous ankylosaurids from North America (e.g. Euoplocephalus tutus, Scolosaurus cutleri).
Somewhat surprisingly, a sexual display function for ankylosaurid cranial ornamentation has not
previously been proposed, despite the fact that sexual display has been proposed as the function of far
less elaborate structures in some theropods and sauropods (see Hone, Naish & Cuthill, 2011 for a
review). Knell et al. (2013) outline five characteristics that can be used to identify potential sexually
selected traits in fossil animals: sexual dimorphism, ontogenetic changes, allometry, phylogenetic
diversity and morphological disparity, and costliness. There are too few specimens to evaluate sexual
dimorphism, ontogenetic changes, or allometry in the Baruungoyot and Nemegt formation
ankylosaurids. However, there is no evidence for sexual dimorphism in the better represented North
American genera Anodontosaurus lambei, Euoplocephalus tutus, and Scolosaurus cutleri (Arbour &

Currie 2013a). Although there is a relatively large sample of juvenile Pinacosaurus grangeri, few adult
specimens are known (Burns et al. 2011), and so assessing ontogenetic or allometric changes in this
genus is also difficult. The bone that forms ankylosaurid cranial ornamentation has a physiological cost,
and would not seem to confer additional protection against predation; although this observation alone
cannot be used to support a sexually-selected interpretation of elaborate ankylosaurid cranial
ornamentation, it does provide support for the suggestion that ankylosaurid cranial ornamentation may
have been at least partly sexually selected. Future studies investigating the morphological disparity of
ankylosaurid cranial ornamentation may provide additional support for this hypothesis.
CONCLUSIONS
A systematic review of ankylosaurid material from the Baruungoyot and Nemegt formations of
Mongolia shows that three species were present in the Baruungoyot Formation and two in the Nemegt
Formation. "Dyoplosaurus giganteus" is a nomen dubium because the holotype lacks any diagnostic
features at the level of genus or species. Tarchia kielanae is here recognized as a valid species based on
the presence of a single autapomorphy, an accessory postorbital ossification surrounded by a distinct
furrow. This feature is also present in the holotype of Minotaurasaurus ramachandrani. As such,
Minotaurasaurus ramachandrani is considered a junior synonym of Tarchia kielanae, which in turn
suggests that INBR21004 may have been collected from the Baruungoyot Formation of Mongolia. A new
genus and species, Zaraapelta nomadis, is characterized by squamosal horns with a distinct bi-layered
texture, and by extensive postocular ornamentation. PIN 3142/250, previously referred to Tarchia
gigantea, is here referred to Saichania chulsanensis. At least two ankylosaurid species were present in
the Nemegt Formation based on tail club morphotypes, but the lack of overlapping postcranial material
precludes the referral of either morphotype to the one named species in the formation, Saichania
chulsanensis. A revised phylogenetic analysis of the Ankylosauridae recovered Tarchia kielanae and
Zaraapelta nomadis as sister taxa, and also showed a close relationship between these species and
Saichania chulsanensis.
ACKNOWLEDGEMENTS
MPC D100/1338 was discovered by Robert Gabbard (Clifton, Colorado) and excavated by Robert
Gabbard, Philip Currie, Mona Marsovsky (Calgary, Alberta) and Albert Miniaci (Fort Lauderdale, Florida).
C. Coy (UALVP) prepared MPC D100/1338, and J. Tansey prepared the illustrations used in Figures 8, 9,
and 10. Many thanks to the following people who provided access to collections and assistance at their
respective institutions: M. Norell and C. Mehling (AMNH), Xu X. and Zheng F. (IVPP), K. Tsogtbaatar and
T. Chinzorig (MPC), T. Tumanova (PIN), and M. Borsuk-Białynicka (ZPAL). Thanks also to Y.-N. Lee and the
Korea-Mongolia International Dinosaur Project 2010 field crew. This manuscript was improved by
discussions with J. Acorn, M. Burns, E. Koppelhus, and A. Murray. Funding was provided to VMA by a
National Sciences and Engineering Research Council Canada Graduate Scholarship – Doctoral and
Michael Smith Foreign Study Supplement, Alberta Ingenuity Studentship, the Dinosaur Research
Institute, the Korea-Mongolia International Dinosaur Project, the University of Alberta China Institute,
and the University of Alberta Graduate Students Association. Comments from S. Maidment, M. Loewen,
and two anonymous reviewers helped improve this manuscript, and the reviewers are thanked for their
time.

LITERATURE CITED
Alicea J, Loewen M. 2013. New Minotaurasaurus material from the Djodokta Formation establishes new
taxonomic and stratigraphic criteria for the taxon. Journal of Vertebrate Paleontology, Program and
Abstracts 2013, p.76.
Arbour VM, Currie PJ. 2013a. Euoplocephalus tutus and the diversity of ankylosaurid dinosaurs in the
Late Cretaceous of Alberta, Canada, and Montana, USA. PLOS ONE 8:e62421.
Arbour VM, Currie PJ. 2013b. The taxonomic identity of a nearly complete ankylosaurid dinosaur
skeleton from the Gobi Desert of Mongolia. Cretaceous Research 46:24-30.
Arbour VM, Burns ME, Sissons RL. 2009. A redescription of the ankylosaurid dinosaur Dyoplosaurus
acutosquameus Parks, 1924 (Ornithischia: Ankylosauria) and a revision of the genus. Journal of
Vertebrate Paleontology 29:1117-1135.
Arbour VM, Burns ME, Currie PJ. 2011. A review of pelvic shield morphology in ankylosaurs (Dinosauria:
Ornithischia). Journal of Paleontology 85:298-302.
Arbour VM, Lech-Hernes NL, Guldberg TE, Hurum JH, Currie PJ. 2013. An ankylosaurid dinosaur from
Mongolia with in situ armour and keratinous scale impressions. Acta Palaeontologica Polonica.
Benton MJ. 2000. Conventions in Russian and Mongolian palaeontological literature. In: Benton MJ,
Shishkin MA, Unwin DM, Kurochkin EN, eds. The Age of Dinosaurs in Russia and Mongolia. Cambridge:
Cambridge University Press, xvi-xxxix.
Blows WT. 2001. Dermal armor of the polacanthine dinosaurs. In: Carpenter K, ed. The Armored
Dinosaurs. Bloomington: Indiana University Press, 363-385.
Brown B. 1908. The Ankylosauridae, a new family of armored dinosaurs from the Upper Cretaceous.
Bulleting of the American Museum of Natural History 24:187-201.
Burns ME, Currie PJ, Sissons RL, Arbour VM. 2011. Juvenile specimens of Pinacosaurus grangeri Gilmore,
1933 (Ornithischia: Ankylosauria) from the Late Cretaceous of China, with comments on the specific
taxonomy of Pinacosaurus. Cretaceous Research 32:174-186.
Carpenter K. 2001. Phylogenetic analysis of the Ankylosauria. In: Carpenter K, ed. The Armored
Dinosaurs. Bloomington: Indiana University Press, 455-483.
Carpenter K. 2004. Redescription of Ankylosaurus magniventris Brown 1908 (Ankylosauridae) from the
Upper Cretaceous of the Western Interior of North America. Canadian Journal of Earth Sciences 41:961-
986.

Carpenter K, Bartlett J, Bird J, Barrick R. 2008. Ankylosaurs from the Price River Quarries, Cedar
Mountain Formation (Lower Cretaceous), east-central Utah. Journal of Vertebrate Paleontology
28:1089-1101.
Carpenter K, Hayashi S, Kobayashi Y, Maryańska T, Barsbold R, Sato K, Obata I. 2011. Saichania
chulsanensis (Ornithischia, Ankylosauridae) from the Upper Cretaceous of Mongolia. Palaeontographica
Abt A 294:1-61.
Coombs WP, Jr. 1971. The Ankylosauria. PhD dissertation, Columbia University, New York, 487 p.
Coombs WP, Jr. 1978. The families of the ornithischian dinosaur Order Ankylosauria. Palaeontology
21:143-170.
Dalton R. 2009. Paper sparks fossil fury: paleontologists criticize publication of specimen with
questionable origin. Nature News doi:10.1038/news.2009.60.
Gallagher WB, Tumanova TA, Dodson P, Axel L. 1998. CT scanning Asian ankylosaurs: paleopathology in
a Tarchia skull. Journal of Vertebrate Paleontology 18:44A-45A.
Galton PM. 1978. Fabrosauridae, the basal family of ornithischian dinosaurs (Reptilia: Ornithischia).
Palaeontologische Zeitschrift 52:138-159.
Gilmore CW. 1923. A new species of Corythosaurus, with notes on other Belly River Dinosauria.
Canadian Field-Naturalist 37:46-52.
Gilmore CW. 1930. On dinosaurian reptiles from the Two Medicine Formation of Montana. Proceedings
of the United States National Museum 77:1-39.
Gilmore CW. 1933. Two new dinosaurian reptiles from Mongolia with notes on some fragmentary
specimens. American Museum Novitates 679:1-20.
Godefroit P, Pereda-Suberbiola X, Li H, Dong Z. 1999. A new species of the ankylosaurid dinosaur
Pinacosaurus from the Late Cretaceous of Inner Mongolia (P.R. China). Bulletin de l'Institut Royal des
Sciences Naturelles de Belgique, Sciences de la Terre 69(suppl.):17-366.
Goloboff P, Farris S, Nixon K. 2008. TNT (Tree analysis using New Technology) ver. 1.1. Published by the
authors, Tucumán, Argentina.
Hill RV, Witmer LW, Norell MA. 2003. A new specimen of Pinacosaurus grangeri (Dinosauria:
Ornithischia) from the Late Cretaceous of Mongolia: ontogeny and phylogeny of ankylosaurs. American
Museum Novitates 3395:1-29.

Hone DWE, Naish D, Cuthill IC. 2011. Does mutual sexual selection explain the evolution of head crests
in pterosaurs and dinosaurs? Lethaia 45:139-156.
Jerzykiewicz T. 2000. Lithostratigraphy and sedimentary settings of the Cretaceous dinosaur beds of
Mongolia. In: Benton MJ, Shishkin MA, Unwin DM, Kurochkin EN, eds. The Age of Dinosaurs in Russia
and Mongolia. Cambridge: Cambridge University Press, 279-296.
Kirkland JI. 1998. A polacanthine ankylosaur (Ornithischia: Dinosauria) from the Early Cretaceous
(Barremian) of eastern Utah. New Mexico Museum of Natural History and Science Bulletin 14:271-281.
Knell RJ, Naish D, Tomkins JL, Hone DWE. 2013. Sexual selection in prehistoric animals: detection and
implications. Trends in Ecology and Evolution 28:38-47.
Lambe LM. 1902. New genera and species from the Belly River Series (mid-Cretaceous). Geological
Survey of Canada Contributions to Canadian Palaeontology 3:25-81.
Lambe LM. 1919. Description of a new genus and species (Panoplosaurus mirus) of an armoured
dinosaur from the Belly River Beds of Alberta. Transactions of the Royal Society of Canada, series 3
13:39-50.
Lee Y-N. 1996. A new nodosaurid ankylosaur (Dinosauria: Ornithischia) from the Paw Paw Formation
(late Albian) of Texas. Journal of Vertebrate Paleontology 16:232-245.
Longrich NR, Currie PJ, Dong Z-H. 2010. A new oviraptorid (Dinosauria: Theropoda) from the Upper
Cretaceous of Bayan Mandahu, Inner Mongolia. Palaeontology 53:945-960.
Maddison WP, Maddison DR. 2011. Mesquite: a modular system for evolutionary analysis, ver. 2.72.
http://mesquiteproject.org.
Maidment SCR, Porro LB. 2010. Homology of the palpebral and origin of the supraorbital ossifications in
ornithischian dinosaurs. Lethaia 43:95-111.
Maleev EA. 1952. [A new ankylosaur from the Upper Cretaceous of Mongolia.] Doklady Akademii Nauk,
SSSR 87:273-276. [In Russian; translation by T. and F. Jeletsky, 1956]
Maleev EA. 1956. [Armored dinosaurs of the Upper Cretaceous of Mongolia, Family Ankylosauridae].
Trudy Paleontol. Inst. Akademiia nauk SSSR 62:51-91. [In Russian; translation by R. Welch]
Mallon JC, Anderson JS. 2013. Skull ecomorphology of megaherbivorous dinosaurs from the Dinosaur
Park Formation (Upper Campanian) of Alberta, Canada. PLOS ONE 8:367172.
Maryańska T. 1977. Ankylosauridae (Dinosauria) from Mongolia. Palaeontologia Polonica 37:85-151.

Maryańska T, Osmolska H. 1974. Pachycephalosauria, a new suborder of ornithischian dinosaurs.
Palaeontologia Polonica 30:45-102.
Maryańska T, Osmolska H. 1975. Protoceratopsidae (Dinosauria) of Asia. Palaeontologica Polonica
33:133-181.
Miles CA, Miles CJ. 2009. Skull of Minotaurasaurus ramachandrani, a new Cretaceous ankylosaur from
the Gobi Desert. Current Science 96:65-70.
Miyashita T, Arbour VM, Witmer LM, Currie PJ. 2011. The internal cranial morphology of an armoured
dinosaur Euoplocephalus corroborated by X-ray computed tomographic reconstruction. Journal of
Anatomy 219:661-675.
Nopcsa F. 1915. Die dinosaurier der Siebenbürgischen Landesteile Ungarns [The dinosaurs of the
Transylvanian province in Hungary]. Mitteilungen Jahrbuch Ungarische Geologische Reichsanstalt 23:1-
26. [In German; translation by DB Weishampel]
Nopcsa F. 1928. Palaeontological notes on reptiles. Geologica Hungarica, Series Palaeontologica 1:1-84.
Nopcsa F. 1929. Dinosaurier reste aus Siebenbürgen. V. Geologica Hungarica Series Palaeontologica 4:1-
76.
Norman DB. 1984. A systematic reappraisal of the reptile order Ornithischia. In: Reif W-E, Westphal F
eds. 3rd Symposium on Mesozoic Terrestrial Ecosystems, Short Papers. Tübingen, Germany: Attempto
Verlag, 157-162.
Osborn HF. 1923. Two Lower Cretaceous dinosaurs from Mongolia. American Museum Novitates 95:1-
10.
Owen R. 1842. Report on British fossil reptiles. Reports of the British Association for the Advancement of
Science 11:60-204.
Owen R. 1861. A monograph of a fossil dinosaur (Scelidosaurus harrisonii, Owen) of the Lower Lias, part
I. Monographs on the British Fossil Reptilia from the Oolitic Formations 1:1-14.
Pang Q, Cheng Z. 1998. A new ankylosaur of Late Cretaceous from Tianzhen, Shanxi. Progress in Natural
Science 8:326-334.
Parish JC, Barrett PM. 2004. A reappraisal of the ornithischian dinosaur Amtosaurus magnus Kurzanov
and Tumanova 1978, with comments on the status of A. archibaldi Averianov 2002. Canadian Journal of
Earth Sciences 41:299-306.

Parks WA. 1924. Dyoplosaurus acutosquameus, a new genus and species of armored dinosaur, and
notes on a skeleton of Prosaurolophus maximus. University of Toronto Studies Geological Series 18:1-35.
Ryan MJ, Evans DC. 2005. Ornithischian dinosaurs. In: Currie PJ, Koppelhus EB, eds. Dinosaur Provincial
Park: a spectacular ancient ecosystem revealed. Bloomington: Indiana University Press, 312-348.
Seeley HG. 1887. On the classification of the fossil animals commonly named Dinosauria. Proceedings of
the Royal Society of London 43:165-171.
Sternberg CM. 1929. A toothless armoured dinosaur from the Upper Cretaceous of Alberta. Canada
Department of Mines Geological Survey Bulletin (Geological Series) 54:28-33.
Sullivan RM. 1999. Nodocephalosaurus kirtlandensis, gen. et sp nov., a new ankylosaurid dinosaur
(Ornithischia: Ankylosauria) from the Upper Cretaceous Kirtland Formation (Upper Campanian), San
Juan Basin, New Mexico. Journal of Vertebrate Paleontology 19:126-139.
Thompson RS, Parish JC, Maidment SCR, Barrett PM. 2012. Phylogeny of the ankylosaurian dinosaurs
(Ornithischia: Thyreophora). Journal of Systematic Palaeontology 10:301-312.
Tumanova TA. 1977. New data on the ankylosaur Tarchia gigantea. Paleontological Zhurnal 4:92-100.
Tumanova TA. 1983. [The first ankylosaur from the Lower Cretaceous of Mongolia.] Trudy Sovmestnoi
Sovestsko-Mongol’skoi Paleontologicheskoi Expeditsii 24:110-120. [In Russian, translation by R. Welch]
Tumanova TA. 1993. [A new armored dinosaur from south-eastern Gobi]. Paleontologicheskii Zhurnal
27:92-98. [In Russian]
Vickaryous MK, Russell AP. 2003. A redescription of the skull of Euoplocephalus tutus (Archosauria:
Ornithischia): a foundation for comparative and systematic studies of ankylosaurian dinosaurs.
Zoological Journal of the Linnean Society 137:157-186.
Vickaryous MK, Russell AP, Currie PJ. 2001a. Cranial ornamentation of ankylosaurs (Ornithischia:
Thyreophora): reappraisal of developmental hypotheses. In: Carpenter K, ed. The Armored Dinosaurs.
Bloomington: Indiana University Press, 318-340.
Vickaryous MK, Russell AP, Currie PJ, Zhao X-J. 2001b. A new ankylosaurid (Dinosauria: Ankylosauria)
from the Lower Cretaceous of China, with comments on ankylosaurian relationships. Canadian Journal
of Earth Sciences 38:1767-1780.
Vickaryous MK, Maryańska T, Weishampel DB. 2004. Ankylosauria. In: Weishampel DB, Dodson P,
Osmolska H, eds. The Dinosauria, 2nd Edition. Berkeley: University of California Press, 363-392.

Weishampel DB, Barrett PM, Coria RA, Loeuff JL, Xu X, Zhao X, Sahni A, Gomani EMP, Noto CR. 2004.
Dinosaur Distribution. In: Weishampel DB, Dodson P, Osmolska H, eds. The Dinosauria, 2nd Edition.
Berkeley: University of California Press, 517-606.
Witmer LM, Ridgely RC. 2008. The paranasal air sinuses of predatory and armored dinosaurs
(Archosauria: Theropoda and Ankylosauria) and their contribution to cephalic structure. Anatomical
Record 291:1362-1388.
(Character statements are found after the figures, and the .nex file is uploaded separately.)

FIGURES
Figure 1. Maps of Mongolia and Omnogovi showing locations of specimen localities.

Figure 2. Ankylosaurid postcranial elements from the Nemegt Formation. A-E, selected elements of PIN 551-9,
holotype of Dyoplosaurus giganteus: (A) pedal phalanges and ungual, (B) free caudal vertebra, (C) posterior free
caudal vertebra, (D) tail club knob, and (E) distal portion of tail club handle. (F) Portion of tail club of ZPAL MgD
I/113, previously referred to Tarchia gigantea (photograph of cast UALVP 47948). The morphology of the handle
vertebrae of PIN 551-9 and ZPAL MgD I/113 differ, as the angle formed by the neural spine in dorsal view is
considerably more acute in PIN 551-9 compared to ZPAL MgD I/113. See list in text for an explanation of
anatomical abbreviations.

Figure 3. Comparison of the supraorbital, postorbital, and squamosal region of three Mongolian ankylosaurid
specimens, in dorsal view. A) PIN 3142/250, Saichania chulsanensis; previously referred to Tarchia gigantea. B)
ZPAL MgD I/111, holotype of Tarchia kielanae (modified from Maryańska 1977). C) INBR 21004, holotype of
Minotaurasaurus ramachandrani (photograph of cast UALVP 49402). INBR 21004 has an unusual accessory
postorbital ossifiation and postorbital furrow that is found in only one other specimen, ZPAL MgD I/111. As such,
Minotaurasaurus ramachandrani should be considered a junior synonym of Tarchia kielanae. PIN 3142/250,
previously referred to Tarchia gigantea, lacks an accessory postorbital ossification and postorbital furrow, and
therefore is not referable to Tarchia; it is here referred to Saichania chulsanensis. See list in text for an explanation
of anatomical abbreviations.

Figure 4. Ankylosaurid skulls in dorsal view, showing differences in cranial ornamentation patterns. (A) AMNH
6523, holotype of Pinacosaurus grangeri. (B) ZPAL MgD II/1, juvenile Pinacosaurus grangeri. (C) IMM 96BM3/1,
holotype of Pinacosaurus mephistocephalus (drawn from Godefroit et al. 1999). (D) MPC 100/151, holotype of
Saichania chulsanensis. (E) PIN 3142/250, Saichania chulsanensis referred specimen. (F) INBR21004, Tarchia
kielanae. (G) MPC D100/1388, holotype of Zaraapelta nomadis, gen. et sp. nov. Skulls are all scaled to the same
anteroposterior length from the anterior end of the skull to the posterior edge of the nuchal crest. See list in text
for an explanation of anatomical abbreviations.

Figure 5. Mongolian ankylosaurid skulls in
left lateral view. A) PIN 3142/250,
Saichania chulsanensis, previously
referred to Tarchia gigantea. B) MPC
100/151, holotype of Saichania
chulsanensis (photograph of cast
mounted with MPC 100/1305, right side
horizontally flipped). C) MPC D100/1388,
holotype of Zaraapelta nomadis, gen. et
sp. nov. D) INBR21004, Tarchia kielanae
(photograph of cast UALVP 49402). E)
AMNH 6523, holotype of Pinacosaurus
grangeri. F) ZPAL MgD II/1, juvenile
Pinacosaurus grangeri. See list in text for
an explanation of anatomical
abbreviations.

Figure 6. Crania of MPC 100/151, holotype of Saichania chulsanensis (A) and INBR21004, Tarchia kielanae (B) in
anterior view. INBR21004 represented by cast specimen UALVP 49402. See list in text for an explanation of
anatomical abbreviations.
Figure 7. Mandibles of Mongolian ankylosaurids in lateral view. A) MPC 100/151, holotype of Saichania
chulsanensis referred specimen, right mandible horizontally mirrored. B) PIN 3140/250, Saichania chulsanensis
(previously referred to Tarchia gigantea) left mandible. C) INBR 21004, Tarchia kielanae left mandible. See list in
text for an explanation of anatomical abbreviations.

Figure 8. MPC D100/1338, holotype of Zaraapelta nomadis, gen. et sp. nov., photograph and interpretive drawing
in lateral view. See list in text for an explanation of anatomical abbreviations.

Figure 9. MPC D100/1338, holotype of Zaraapelta nomadis, gen. et sp. nov., photograph and interpretive drawing
in dorsal view. See list in text for an explanation of anatomical abbreviations.

Figure 10. MPC D100/1338, holotype of Zaraapelta nomadis, gen. et sp. nov., photograph and interpretive drawing
in ventral view. See list in text for an explanation of anatomical abbreviations.

Figure 11. MPC D100/1338, holotype of Zaraapelta nomadis, gen. et sp. nov. A) Skull in posterior view. B) Skull in
anterior view showing cross-section through rostrum. C) Detail of braincase, oblique anteroventral view. D) Detail
of braincase, oblique anterolateral view. See list in text for an explanation of anatomical abbreviations.

Figure 12. Strict consensus, 50% majority-rule, and maximum agreement subtrees recovered from the cladistic
analysis using the traditional search option in TNT, including bootstrap values (1000 replicates) and Bremer
support values. Dyoplosaurus acutosquameus, Nodocephalosaurus kirtlandensis, and Tianzhenosaurus youngi were
pruned from the maximum agreement subtree.

Appendix: Character Statements
General notes:
Codings for Tarchia kielanae are the same as those for Minotaurasaurus ramachandrani in Arbour &
Currie (2013), plus additional new characters in this analysis. Postcranial character codings from Tarchia
gigantea were not transferred to Tarchia kielanae or Saichania chulsanensis, because no postcranial
specimens can be definitely attributed to these species at this time.
Character addition and removal:
Characters 147-160 are new to this analysis. The following characters from Thompson et al. (2012) and
Arbour and Currie (2013) were removed from this analysis:
Characters 9-12: These characters concern the morphology of the airway and paranasal sinuses. Witmer
and Ridgely (2008) have shown that the respiratory passage and lateral sinus in Euoplocephalus both
represent a complex looping airway. These characters have been replaced by the new character 148.
Character 17: It is not always possible to determine which specimens represent adult individuals, and so
it is not possible to assess whether or not fusion of cranial elements represents ontogenetic vs.
taxonomic differences.
Characters 27-28: These characters concern the palpebral, which is not present in ankylosaurs (see
Maidment and Porro 2010).
Character 31: No ankylosaurids have domed parietals, and this character is largely relevant to
nodosaurids.
Characters 52 and 53: The relative depth of the distal ends of the paroccipital processes, and the relative
thickness of bone at the dorsal margin of the foramen magnum, are difficult to quantify and code
consistently.
Character 55: This character needs to be reassessed, as it is not immediately clear how states 0
(transversely convex) and 1 (medial depression) differ.
Character 60: Weak vs. strong endocranial flexure is difficult to assess.
Characters 71 and 72: The ventral margin of the dentary is straight in all ankylosaurids where it is
preserved. Character 72, the shape of the alveolar margin, could not be assessed because it was not
clear in which view the character referred to.
Character 79: No ankylosaurids have completely random distributions of cranial caputegulae, but there
is always some amount of asymmetry. This character did not provide a cut-off for what constitutes
random vs. symmetrical.
Character 85: This character described the shape of the squamosal horn, which is replaced by an
expanded character further in the list.

Character 106: This character described the number of sacral vertebrae. Because ankylosaurs
incorporate numerous dorsal and caudal vertebrae into the sacral rod, it is unclear if this character
referred only to 'true' sacral vertebrae, or all dorsosacrals, sacrals, and caudosacrals. It is also not clear
how this number is affected by ontogeny or size.
Character 109: The length of the transverse processes varies along the vertebral column, and although
this character referred to the proximal caudals, it was not clear just how proximal they had to be in
order to be included.
Characters 143-146: These characters describe features of the pubis, which is markedly reduced or
potentially absent in most ankylosaurids, and preserved in very few specimens.
Characters 160-161: These characters described osteoderm morphology. Character 160 applied only to
Stegosaurus. Character 161 (lateral rows of osteoderms on dorsal aspect of trunk) was redundant with
character 159 (parasaggital rows of keeled osteoderms on dorsal aspect of trunk).
Character 163: The presence of 'quarter rings' in ankylosaurids is based on a misinterpretation of the
cervical half rings in Ankylosaurus. No ankylosaurids have quarter rings – the rings always form a semi-
circle over the cervical vertebrae.
Characters 164-165: Pectoral spikes are more relevant to nodosaurids and early ankylosaurs or
ankylosaurids.
Character 167: This character refers to the form of ossicles in sacral armour; ossicles are not well
defined, and so this character is difficult to assess.
Character 168: This character refers to the size of the lateral trunk plates; 'large' and 'small' were
undefined, and 'large' was also associated with 'hollow', suggesting that this character should be broken
into two separate characters.
Character 169: This character is difficult to assess.
Characters:
1. Antorbital fenestra: present (0); absent (1). (Sereno 1999: 8; Thompson et al. 2012: 1)
2. Modified: Lateral temporal fenestra, visible in lateral view: visible (0); not visible (1). (Carpenter et
al. 1998 : 6 ; Thompson et al. 2012: 2).
3. Supratemporal fenestra: open (0); closed (1). (Lee 1996: 2; Thompson et al. 2012: 3).
4. Skull dimensions, including ornamentation: longer than wide (0); as wide, or wider than long (1).
(Carpenter et al. 1998: 1; Thompson et al. 2012: 4).
5. Modified: Width of the posterior margin of the skull (including squamosal horns) relative to the
maximum width across the orbits: greater or equal (0); less (1). (Vickaryous et al. 2004: 6;
Thompson et al. 2012: 5).
6. Size of occiput: higher than wide (0); wider than high (1). (Lee 1996: 1; Thompson et al. 2012: 6).

7. Modified: External nares, defined as the outermost rim of the nasal vestibule, opening faces:
laterally (0); anterolaterally (1); anteriorly (2). (Carpenter et al. 1998: 10; Thompson et al. 2012: 7).
8. External nares, visible in dorsal view: visible (0); hidden (1). (Thompson et al. 2012: 8)
9. Orbits, angle of orbital axis: <40º (0); >40º (1). (Thompson et al. 2012: 13).
10. Antorbital region of the dorsal skull surface: flat (0); domed (1). (Sereno 1999: 99; Thompson et al.
2012: 14).
11. Development of the postocular shelf: not developed (0); completely separating orbit from
temporal space (1). (Sereno 1999: 104; Thompson et al. 2012: 15).
Updated codings: Tianzhenosaurus from 1 to ?. Notes: Cannot be determined from Pang and
Cheng (1998).
12. Gap between palate and braincase: open (0); closed by a dorsal projection of the pterygoid (1).
(Sereno 1999: 61; Thompson et al. 2012: 16).
13. Dimensions of premaxillary palate: longer than wide (0); wider than long (1). (Vickaryous et al.
2001: 13; Thompson et al. 2012: 18).
14. Shape of the premaxillary palate: sub-triangular (0); sub-quadrate (1); sub-oval (2). (Sereno 1999:
80; Thompson et al. 2012: 19).
15. ‘V’ or ‘U’-shaped median indentation of the anterior margin of the premaxilla: absent (0); present
(1). (Sereno 1999: 91; Thompson et al. 2012: 20).
16. Caudoventral extension of premaxillary tomium in lateral view: ends anteriorly to the maxillary
teeth (0); obscures anteriormost maxillary teeth (1). (Sereno 1999: 100; Thompson et al. 2012:
21).
17. Bone bordering anterior margin of internal nares: premaxilla (0); maxilla (1). (Thompson et al.
2012: 22).
18. Shape of the ventral margin of premaxillary tomium in lateral view: flat (0); convex (1); concave
(2). (Vickaryous et al. 2004: 12; Thompson et al. 2012: 24).
19. Shape of the maxillary tooth row: straight (0); medially convex (1). (Vickaryous et al. 2001: 18;
Parish 2005: 20).
20. Maxillary tooth row position: lateral margin of skull (0); inset (1). (Lee 1996: 4; Thompson et al.
2012: 25).
21. Modified: Distance between posteriormost extent of maxillary tooth rows relative to the width of
the premaxillary beak: wider (0); narrower (1). [The width of the premaxillary beak is measured
where the lateral edges of the beak are most parallel, which is usually close to the posterior of the
premaxilla.] (Sereno 1999: 102; Thompson et al. 2012: 26).
22. Anterior and posterior supraorbitals (recognisable by distinct regions of ornamentation above the
orbit): absent (0); present (1). (Sereno 1999: 13; Thompson et al. 2012: 29).
23. Form of supraorbital ornamentation: boss-like, rounded laterally (0); sharp lateral rim, forming a
ridge (1). (Vickaryous et al. 2001: 5; Thompson et al. 2012: 30).
24. Proportions of jugal orbital ramus: depth greater than transverse breadth (0); transverse breadth
greater than depth (1). (Sereno 1999: 1; Thompson et al. 2012: 32).

25. Shape of quadrate in lateral aspect: curved (anteriorly convex, posteriorly concave) (0); straight
(1). (Vickaryous et al. 2001: 38; Thompson et al. 2012: 33).
26. Inclination of quadrate in lateral aspect: near vertical (0); almost 45º rostrolaterally (1). (Lee 1996:
10; Thompson et al. 2012: 34).
27. Form of the anterior surface of the quadrate: transversely concave (0); not concave (1). (Lee 1996:
12; Thompson et al. 2012: 35).
28. Ventral projection of the mandibular process of the quadrate in lateral view: projects beyond the
quadratojugal ornamentation (0); hidden by quadratojugal ornamentation (1). (Vickaryous et al.
2004 : 40; Thompson et al. 2012: 36).
29. Form of quadrate mandibular extremity: symmetrical (0); medial condyle larger than lateral
condyle (1). (Sereno 1999: 10; Thompson et al. 2012: 37).
30. Inclination of the articular surface of the quadrate condyle in posterior view: horizontal (0);
ventromedially inclined at approximately 45° to horizontal (1). (Sereno 1999: 14; Thompson et al.
2012: 38).
31. Lateral ramus of the quadrate: present (0); absent (1). (Sereno 1999: 15; Thompson et al. 2012:
39).
32. Dorsoventral depth of the pterygoid process of the quadrate: deep (0); shallow (1). (Lee 1996: 7;
Sereno 1999: 60; Thompson et al. 2012: 40).
33. Contact between paroccipital process and quadrate: sutural (0); fused (1). (Carpenter et al. 1998:
13; Thompson et al. 2012: 42).
34. Contact between pterygoids: pterygoids separate caudomedially, forming an interpterygoid
vacuity (0); pterygoids joined medially forming a pterygoid shield (1). (Thompson et al. 2012: 42).
35. Direction of the pterygoid flange: anterolateral (0); anterior/parasagittal (1). (Vickaryous et al.
2001: 29; Thompson et al. 2012: 43).
36. Contact between basipterygoid processes and pterygoid: sutural (0); fused (1). (Vickaryous et al.
2001: 30; Thompson et al. 2012: 44).
37. Position of ventral margin of the pterygovomerine keel relative to alveolar ridge: dorsal (0); level
(1). (Sereno 1999: 59; Thompson et al. 2012: 45).
38. Dorsal extent of median vomer lamina: does not meet skull roof (0); meets skull roof (1). (Lee
1996: 14; Thompson et al. 2012: 46).
39. Pterygoid foramen: absent (0); present (1). (Hill et al. 2003: 21, Thompson et al. 2012: 47).
40. Position of posterior margin of pterygoid body relative to the anterior margin of the quadrate
condyle: anteriorly positioned (0); in transverse alignment (1). (Vickaryous et al. 2004: 28;
Thompson et al. 2012: 48).
41. Caudoventral secondary palate: absent (0); present (1). (Vickaryous et al. 2004: 21, Thompson et
al. 2012: 49)
42. Posterior palatal foramen: absent (0); present (1). (Lee 1996: 17; Thompson et al. 2012: 50).
43. Direction of paroccipital process extension: caudolateral (0); lateral (1). (Carpenter et al. 1998: 11;
Vickaryous et al. 2004: 33 ; Thompson et al. 2012: 51).

44. Bones forming the occipital condyle: basioccipital and exoccipital (0); basioccipital only (1). (Lee
1996: 9; Thompson et al. 2012: 54).
45. Length of basisphenoid relative to the basioccipital: longer (0); shorter or equal (1). (Sereno 1999:
12; Thompson et al. 2012: 56).
46. Form of basisphenoidal tuberosities: medially separated rounded rugose stubs (0); continuous
transverse rugose ridge (1). (Vickaryous et al. 2001: 32; Thompson et al. 2012: 57).
47. Size of basipterygoid processes: twice as long as wide or over (0); less than twice as long as wide
(1). (Thompson et al. 2012: 58).
48. Form of the cranial nerve foramina IX-XII: separate foramina (0); single foramen shared with the
jugular vein (1). (Thompson et al. 2012: 59).
49. Direction of occipital condyle: posterior (0); posteroventral (1). (Vickaryous et al. 2004: 36;
Thompson et al. 2012: 61).
50. Direction of the foramen magnum: posterior (0); posteroventral (1). (Vickaryous et al. 2004: 37;
Thompson et al. 2012: 62).
51. Premaxillary teeth: present (0); absent (1). (Sereno 1999: 18; Thompson et al. 2012: 63).
52. Cingula on maxillary and/or dentary teeth: absent (0); present (1). (Carpenter et al. 1998: 21;
Thompson et al. 2012: 64).
53. Maxillary and/or dentary tooth crown shape: ≥13 denticles, tooth crown pointed (0); <13
denticles, tooth crown rounded (1). (Thompson et al. 2012: 65).
54. Number of dentary teeth: <25 (0); ≥25 (1). (Thompson et al. 2012: 66).
55. Position of mandible articulation relative to mandibular adductor fossa: posterior (0);
posteromedial (1). (Sereno 1999: 64; Thompson et al. 2012: 67).
56. Mandibular fenestra: present (0); absent (1). (Thompson et al. 2012: 68).
57. Depth of the dentary symphysial ramus relative to half the maximum depth of the mandibular
ramus in lateral view: deeper (0); shallower (1). (Sereno 1999: 17; Thompson et al. 2012: 69).
58. Shape of dorsal margin of the dentary in lateral view: straight (0); sinuous (1). (Sereno 1999: 4;
Thompson et al. 2012: 70).
59. Development of the coronoid process: not developed (0); distinct (1). (Sereno 1999: 108;
Thompson et al. 2012: 73).
60. Position of glenoid for quadrate relative to mandibular axis: medially offset (0); in line (1). (after
Carpenter et al. 1999; Thompson et al. 2012: 74).
61. Size and projection of the retroarticular process: small with no dorsal projection (0); well
developed with a dorsal projection (1). (Thompson et al. 2012: 75).
62. Size of predentary ventral process: distinct, prong shaped process (0); rudimentary eminence (1).
(Sereno 1999: 66; Thompson et al. 2012: 76).
63. Ornamentation, defined as sculpturing of skull bones or addition of osteoderms: absent (0);
present (1). (Sereno 1999: 63; Thompson et al. 2012: 77).

64. Modified: Frontonasal and/or frontoparietal cranial ornamentation: rugose, not differentiated into
discrete polygons (caputegulae) (0), differentiated into discrete polygons (caputegulae) (1). (after
Carpenter et al. 1999; Thompson et al. 2012: 78).
65. A single large medial polygon of ornamentation in the parietal region: absent (0); present (1).
(Thompson et al. 2012: 80)
66. Modified: Median nasal caputegulum (located posterior to the supranarial ornamentation, on the
midline of the skull): absent (0), present, hexagonal (1), present, triangular (1). (Vickaryous et al.
2004: 9; Thompson et al. 2012: 81).
67. Modified: Frontonasal caputegulum relief: concave to flat (low relief) (0), bulbous (high relief) (1).
(after Sullivan 1999 ; Thompson et al. 2012: 82)
68. Modified: Projection of squamosal horns relative to the posterior margin of the dorsal surface of
the skull: horns do not project past posterior margin of skull in dorsal view (0), horns project past
posterior margin of skull in dorsal view (1). (Thompson et al. 2012: 83).
69. Modified: Squamosal horn: absent (0); present (1). (Lee 1996: 18; Thompson et al. 2012: 84.
70. Quadratojugal ‘horn’: absent (0); present (1). (after Carpenter et al. 1999: Thompson et al. 2012:
86).
71. Modified: Shape of quadratojugal horn in dorsal view: U-shaped, with round distal edge (0),
triangular, with pointed distal edge (1). (Thompson et al. 2012: 85)
72. Modified: nuchal ornamentation (at posterior margin of skull roof): absent (0); present (1).
(Vickaryous et al. 2004: 11; Thompson et al. 2012: 88).
73. Posterior projection of the nuchal shelf: does not obscure occiput in dorsal view (0); obscures
occiput in dorsal view (1). (Vickaryous et al. 2004: 12; Thompson et al. 2012: 89).
74. Modified: Length of mandibular caputegulum with respect to the length of the mandible: less than
or equal to half the length (0); over three quarters the length (1). (after Carpenter et al. 1999;
Thompson et al. 2012: 90).
75. Mandibular osteoderm: absent (0); present (1). (Thompson et al. 2012: 91)
76. Type of articulation between the atlantal neural arch and intercentrum: open (0); fused in adult
(1). (Sereno 1999: 19; Thompson et al. 2012: 92).
77. Type of contact between the atlantal neural arches: no median contact (0); median contact (1).
(Sereno 1999: 68; Thompson et al. 2012: 93).
78. Contact between atlas and axis: articulated (0); fused (1). (Vickaryous et al. 2004: 46; Thompson et
al. 2012: 94).
79. Dimensions of cervical vertebrae centra: anteroposteriorly longer than transverse width (0);
anteroposteriorly shorter than transverse width (1). (after Kirkland et al. 1998; Thompson et al.
2012: 95).
80. Ratio of maximum neural spine width to height in anterior cervicals: <0.25 (0); ≥0.25 (1). (after
Carpenter et al. 1999; Thompson et al. 2012: 96).
81. Alignment of anterior and posterior faces of cervical centra: aligned (0); anterior face dorsal to
posterior face (1); anterior face ventral to posterior face (2). (Vickaryous et al. 2004: 47;
Thompson et al. 2012: 97).

82. Ratio of anteroposterior [dorsal] centrum length to posterior centrum height: >1.1 (0); <1.1 (1).
(Thompson et al. 2012: 98).
83. Longitudinal keel on ventral surface of dorsal centra: present (0); absent (1). (Parish 2005: 90).
(Thompson et al. 2012: 99)
84. Cross sectional shape of neural canal in posterior dorsals: circular (0) elliptical, with long axis
running dorsoventrally (1). (after Carpenter 1990; Thompson et al. 2012: 100).
85. Shape of the proximal cross-section of the dorsal ribs: triangular (0); ‘L’- or ‘T’-shaped (1).
(Thompson et al. 2012: 101).
86. Attachment of dorsal ribs to posterior dorsal vertebrae: articulated (0); fused (1). (Thompson et al.
2012: 102).
87. Contact between posteriormost dorsal vertebrae: articulated (0); fused to form a presacral rod (1).
(Thompson et al. 2012: 103).
88. Paravertebrae: absent (0); present (1). (Thompson et al. 2012: 104)
89. Longitudinal groove in ventral surface of the sacrum: absent (0); present (1). (Thompson et al.
2012: 105)
90. Ratio of maximum distal width to height of the neural spines of proximal caudals: ≤0.2 (0); >0.2
(1). (after Carpenter 2001; Thompson et al. 2012: 107).
91. Direction of the transverse processes of proximal caudals: craniolaterally projecting (0);
caudolaterally projecting (1); laterally projecting (2). (after Carpenter 2001; Thompson et al. 2012:
108).
92. Persistence of transverse processes down the length of the caudal series: not present beyond the
mid-length of the series (0); present beyond the mid-length of the series (1). (Thompson et al.
2012: 110).
93. Attachment of haemal arches to their respective centra: articulated (0); fused (1). (Thompson et
al. 2012: 111).
94. Shape of distal caudal postzygapophyses: short with a sub-triangular end [wedge-shaped] (0); long
with a rounded end [tongue shaped] (1). (Sereno 1999: 110; Thompson et al. 2012: 112).
95. Extent of pre- and postzygapophyses over their adjacent centra in posterior vertebrae: extend
over less than half the length of the adjacent centrum (0); extend over more than half the length
of the adjacent centrum (1). (Sereno 1999: 109; Thompson et al. 2012:).
96. Shape of the posterior haemal arches: rounded haemal spine in lateral view with no contact
between haemal arches (0); inverted ‘T’-shaped haemal spine in lateral view, with contact
between the ends of adjacent spines (1). (Sereno 1999: 71; Thompson et al. 2012: 114).
97. Ossified tendons in distal region of tail: absent (0); present (1). (Sereno 1999: 97; Thompson et al.
2012: 115).
98. Dimensions of coracoid: longer than wide (0); wider than long or equal width and length (1).
(Thompson et al. 2012: 116).
99. Form of the anterior margin of the coracoid: convex (0); straight (1). (Thompson et al. 2012: 117).
100. Cranioventral process of coracoid: absent (0); present (1). (Thompson et al. 2012: 118).

101. Size of coracoid glenoid relative to scapula glenoid: sub-equal (0); half the size (1). (Sereno 1999:
89; Thompson et al. 2012: 119).
102. Contact between scapula and coracoid: articulated (0); fused (1). (Thompson et al. 2012: 120).
103. Scapula glenoid orientation: ventrolateral (0); ventral (1). (Sereno 1999: 87; Thompson et al. 2012:
121).
104. Ventral process of scapula at the caudoventral margin of glenoid: absent (0); present (1).
(Thompson et al. 2012: 122).
105. Form of the scapula acromion process: not developed or ridge-like along the dorsal border of the
scapula (0) tab-like, perpendicular to scapular blade (1) flange-like and folded over towards the
scapula glenoid (1) ridge terminating in a knob-like eminence (2). (Vickaryous et al. 2004: 52;
Thompson et al. 2012: 123)
106. Orientation of the acromion process of scapula: directed away from the glenoid (0); directed
towards scapula glenoid (1). (after Kirkland 1998; Thompson et al. 2012: 124).
107. Scapulocoracoid buttress: absent (0); present (1). (Thompson et al. 2012: 125).
108. Distal end of scapula shaft: narrow (0); expanded (1). (Sereno 1999: 20; Thompson et al. 2012:
126).
109. Contact between sternal plates: separate (0); fused (1). (Sereno 1999: 112; Vickaryous et al. 2004:
60; Thompson et al. 2012: 127).
110. Separation of humeral head and deltopectoral crest in anterior view: continuous (0); separated by
a distinct notch (1). (Thompson et al. 2012: 128).
111. Separation of humeral head and medial tubercle in anterior view: continuous (0); separated by a
distinct notch (1). (Thompson et al. 2012: 129)
112. Ratio of deltopectoral crest length to humeral length: ≤0.5 (0); >0.5 (1). (Thompson et al. 2012:
130).
113. Orientation of deltopectoral crest projection: lateral (0); anterolateral (1). (Sereno 1999: 113;
Thompson et al. 2012: 131).
114. Shape of the radial condyle of humerus round / proximal end of radius in end-on view: non-
circular (0); circular (1). (Thompson et al. 2012: 132).
115. Ratio of the length of metacarpal V to metacarpal III: ≤0.5 (0); >0.5 (1). (Sereno 1999: 6; Thompson
et al. 2012: 133).
116. Manual digit number: 5 (0); 4 (1); 3 (2). (Thompson et al. 2012: 134).
117. Shape of manual and pedal ungual phalanges: claw shaped (0); hoof shaped (1). (Sereno 1999: 7;
Thompson et al. 2012: 135).
118. Length of the preacetabular process of ilium as a percentage of total ilium length: ≤ 50% (0); > 50
%.( Thompson et al. 2012: 136).
119. Angle of lateral deflection of the preacetabular process of the ilium: 10º–20º (0); 45º (1). (Sereno
1999: 21; Thompson et al. 2012: 137).
120. Orientation of the preacetabular portion of the ilium: near vertical (0); near horizontal (1).
(Kirkland 1998: 45; Thompson et al. 2012: 138).

121. Form of the preacetabular portion of the ilium: straight process (0); pronounced ventral curvature
(1). (Thompson et al. 2012: 139).
122. Lateral exposure of the acetabulum: exposed (0) acetabulum partially obscured as it is partially
encircled by the distal margin of the ilium (1). (Thompson et al. 2012: 140)
123. Perforation of the acetabulum: present, open acetabulum (0); absent, closed acetabulum (1).
(Sereno 1999: 74; Thompson et al. 2012: 141).
124. Postacetabular ilium length, relative to diameter of acetabulum: greater (0); smaller (1). (Sereno
1999: 114; Thompson et al. 2012: 142).
125. Shape of ischium: straight (0); ventrally flexed at mid-length (1). (Kirkland 1998: 37; Thompson et
al. 2012: 147).
126. Shape of the dorsal margin of ischium: straight or concave (0); convex (1). (Sereno 1999: 115;
Thompson et al. 2012: 148).
127. Angle between long axis of femoral head and long axis of shaft: <100º (0); 100º to 120º (1); >120º
(2). (Thompson et al. 2012: 149).
128. Separation of femoral head from greater trochanter: continuous (0); separated by a distinct notch
or change in slope (1). (Thompson et al. 2012: 150).
129. Differentiation of the anterior trochanter of the femur: separated from femoral shaft by a deep
groove laterally and dorsally (0); fused to femoral shaft (1). (Kirkland 1998: 36; Thompson et al.
2012: 151).
130. Oblique ridge on lateral femoral shaft, distal to anterior trochanter: absent (0); present (1).
(Thompson et al. 2012: 152).
131. Form of the fourth trochanter: pendant (0); ridge-like (1). (Sereno 1999: 24; Thompson et al. 2012:
153).
132. Location of the fourth trochanter on the femoral shaft: proximal (0) distal, over half-way down the
femoral shaft (1). (Thompson et al. 2012: 154).
133. Maximum distal width of the tibia, compared to the maximum proximal width: narrower (0);
wider (1). (Sereno 1999: 188; Thompson et al. 2012: 155).
134. Contact between tibia and astragalus: articulated (0); fused, with suture obliterated (1).
(Thompson et al. 2012: 156).
135. Number of pedal digits: 5 (0); 4 (1); 3 (2). (Thompson et al. 2012: 157).
136. Phalangeal number in pedal digit IV: 5 (0); ≤4 (1). (Sereno 1999: 26; Thompson et al. 2012: 158).
137. Parasagittal row of keeled osteoderms situated on the dorsal aspect of the trunk: absent (0);
present (1). (Sereno 1999: 2; Thompson et al. 2012: 159).
138. Number of distinct cervical pectoral bands: none (0); one (1); two (2). (Kirkland 1998: 38;
Thompson et al. 2012: 162).
139. Sacral shield of fused osteoderms: absent (0); present (1). (Kirkland 1998: 42; Thompson et al.
2012: 166).
140. Form of sacral armour: rosettes (0), evenly-sized polygons (1). (Thompson et al. 2012: 167)

141. [Modified] Terminus of tail enveloped by >2 osteoderms, forming tail club knob: absent (0),
present (1) (Kirkland 1998: 44, Thompson et al. 2012: 170).
142. Small (<2 cm diameter), circular osteoderms posterolateral to orbit, along ventral edge of
squamosal horn and/or along dorsal edge of quadratojugal horns: absent (0); present (1). (Arbour
and Currie 2013: 171)
143. Cervical half rings: composed of osteoderms that are either tightly adjacent to one another or
coossified at the edges, forming arc over the cervical region (0), composed of osteoderms and
underlying bony band segments, osteoderms may or may not cossify to the band, forming arc over
the cervical region (1). (Arbour and Currie 2013: 172)
144. Composition of first cervical half ring: first cervical half ring has 4 to 6 primary osteoderms only
(0), first cervical half ring has 4 to 6 primary osteoderms surrounded by small (<2 cm diameter)
circular secondary osteoderms. (Arbour and Currie 2013: 173)
145. Form of caudal osteoderms: dorsoventrally compressed, triangular in dorsal view (0), or low cones
(1). (Arbour and Currie 2013: 174)
146. Tail club knob shape: knob absent (0), major knob osteoderms semicircular in dorsal view (1),
triangular in dorsal view (2). (Arbour and Currie 2013: 175)
147. Tail club knob proportions: knob absent (0), tail club knob length > width (1), length = width (2),
width > length (3). (Arbour and Currie 2013: 176)
148. New character: Shape of respiratory passage: straight or arched (0), with anterior (rostral) and
posterior (caudal) loops (sensu Witmer and Ridgely 2008). [Replaces characters 9-12 in Thompson
et al. (2012) and Arbour & Currie (2013).]
149. New character: Lacrimal incisure (Mediolateral constriction behind the narial osteoderms/at the
prefrontals, giving the skull an hourglass-shaped outline in dorsal view): absent (0) present (1)
150. New character: Domed caputegulae: rounded cones with circular bases (0) pyramidal with sharp
edges (1)
151. New character: Number of internarial caputegulae: none (0), 1 (1), more than 1 (2)
152. New character: Supranarial caputegulae, notch dorsal to nasal vestibule absent (0), present (1).
153. New character: Loreal caputegulum in lateral view: 1 caputegulum (0), more than 1 caputegulum
(1)
154. New character: Lacrimal caputegulum in lateral view: 1 caputegulum (0), more than 1
caputegulum (1)
155. New character: Prefrontal osteoderm: flat with keel (0), sharply pointed and pyramidal (1).
156. New character: Depth of jugal ramus relative to orbit height: jugal height is less than 15% orbit
height (0), jugal height is more than 15% orbit height (1)
157. New character: Supraorbital caputegulae, when viewed dorsally: combine to form continuous
edge (0), have distinct apices (1)
158. New character: Accessory postorbital ossification: absent (0), present (1)
159. New character: Quadratojugal horn: lacks distinct neck at base (0), has distinct neck at base (1).
160. New character: Squamosal horn: base has broad triangular cross-section and overall shape is
pyramidal (0), base is oval in cross-section and overall shape is narrow, tapered cylinder (1)

Literature Cited:
1. Arbour VM, Currie PJ. 2013. Euoplocephalus tutus and the diversity of ankylosaurid dinosaurs in
the Late Cretaceous of Alberta, Canada, and Montana, USA. PLOS ONE 8:e62421.
2. Thompson RS, Parish JC, Maidment SCR, Barrett PM (2012) Phylogeny of the ankylosaurian
dinosaurs (Ornithischia: Thyreophora). Journal of Systematic Palaeontology 10:301-312.
3. Sereno PC (1999) The evolution of dinosaurs. Science 284:2137-2147.
4. Carpenter K, Miles C, Cloward K (1998) Skull of a Jurassic ankylosaur (Dinosauria). Nature
393:782-783.
5. Lee Y-N (1996) A new nodosaurid ankylosaur (Dinosauria: Ornithischia) from the Paw Paw
Formation (Late Albian) of Texas. Journal of Vertebrate Paleontology 16:232-245.
6. Vickaryous MK, Maryańska T, Weishampel DB (2004) Ankylosauria. In: Russell DB, Dodson P,
Osmólska H, editors. The Dinosauria, 2nd Edition. Berkeley: University of California Press, pp.
363–392.
7. Pang Q, Cheng Z (1998) A new ankylosaur of Late Cretaceous from Tianzhen, Shanxi. Progress in
Natural Science 8:326-334.